Final Technical Report

COVER PAGE

Report submitted to: Maritime Security Center, Stevens Institute of Technology

Grant Recipient: Florida Atlantic University, DUNS Number: 004147534

Sub-Award Number: 2102833-01

Project Title: Predictive Port Resilience Tool to Assess Regional Impacts of Hurricanes

The project was conducted as a Maritime Security Center project.

Project Period: Start Date: July 1, 2018 to June 30, 2019

Principle Investigator:

Dr. Manhar Dhanak

Professor & Department Chair,

Department of Ocean and Mechanical Engineering,

Florida Atlantic University, 777 Glades Road, Boca Raton, FL 33431

Email: dhanak@fau.edu; Phone: 561 297 2827

Co-Principal Investigator:

Dr. Evangelos Kaisar

Associate Professor and Director,

Freight Mobility Research Institute,

Florida Atlantic University, 777 Glades Road, Boca Raton, FL 33431

Email: ekaisar@fau.edu; Phone: 561 297 4084

Co-Principal Investigator:

Dr. Scott Parr

Assistant Professor

Department of Civil Engineering

Embry-Riddle Aeronautical University, 600 S. Clyde Morris Blvd.,

Daytona Beach, FL 32114

Email: PARRS1@erau.edu; Phone: (386) 226 – 7530

Co-Principal Investigator:

Dr. Alka Sapat

Professor, School of Public Administration and

Coordinator, Bachelor of Public Safety Administration Program

Florida Atlantic University, 777 Glades Road, Boca Raton, FL 33431

Email: asapat@fau.edu; Phone: 561-297-2330

Report Authors: Manhar Dhanak, Panagiota Goulianou, Evangelos Kaisar, Scott Parr, Alka

Sapat, Hannah Thomas, LT Joshua Villafane, USCG and LT Maxwell Walker, USCG

Date of Report: July 2019

2

TABLE OF CONTENTS

EXECUTIVE SUMMARY .............................................................................................. 4

ABBREVIATIONS AND ACRONYMS......................................................................... 6

1.0 INTRODUCTION .................................................................................................. 7

1.2 Research Goal, Objectives, and Broader Impact .................................................... 8

2.0 LITERATURE REVIEW ..................................................................................... 10

2.1 US Coast Guard Instructions ................................................................................ 10

2.2 Resiliency Assessment ......................................................................................... 10

2.3 Port Modeling ....................................................................................................... 15

3.0 STAKEHOLDER ENGAGEMENT AND CASE STUDY PORTS .................... 17

3.1 Case-Study Ports .................................................................................................. 17

3.1.1 Port of Miami ................................................................................................ 18

3.1.2 Port Everglades .............................................................................................. 18

3.1.3 Port of Jacksonville ....................................................................................... 19

3.1.4 Port Canaveral ............................................................................................... 20

3.1.5 Port of Savannah............................................................................................ 21

3.1.6 Port of Charleston .......................................................................................... 21

4.0 METHODOLOGY ............................................................................................... 22

4.1 Resiliency Definition and Assessment ................................................................. 22

4.2 Hybrid Multimodal Simulation Model Development .......................................... 24

4.3 Landside Model Development in VISSIM ........................................................... 26

4.4 Waterside Model Development ............................................................................ 31

4.6 Integrating the Water and Landside Simulation Models ...................................... 36

4.7 Scenario Development.......................................................................................... 37

5.0 RESULTS ............................................................................................................. 39

5.1 Model Calibration and Validation ........................................................................ 39

5.2 Scenario Results ....................................................................................................... 42

5.3 Regional Impact........................................................................................................ 47

6.0 DISCUSSION AND CONCLUSIONS ................................................................ 49

6.1 Plan for Transition ................................................................................................ 50

7.0 ACKNOWLEDGEMENTS .................................................................................. 51

8.0 REFERENCES ......................................................................................................... 52

3

APPENDIX A – ARRIVAL AND DWELL TIME PROBABILITY DISTRIBUTION

FUNCTIONS ..................................................................................... 57

Containerized Cargo ................................................................................................... 57

Non-Containerized Cargo ........................................................................................... 63

Tanker Vessels............................................................................................................ 69

Passenger Vessels ....................................................................................................... 74

4

EXECUTIVE SUMMARY

The work presented here was carried out as a Maritime Security Center project. A modeling and

simulation-based framework has been developed for predicting the consequences of disruption at

ports in a region due to the passage of a hurricane event. This research seeks to expand

understanding of port resiliency and enhance the knowledge in the development of a stakeholder-

focused tool to improve regional resilience. The tool uses modeling and simulation to predict

consequences of disruptions at regional ports due to a hurricane event, in support of seeking

improvements in the regional preparedness of ports. The effort is based in particular on adapting

archival Nationwide Automatic Identification System (NAIS) and Department of Transportation

(DOT) data for resilience analyses of the coastal ports affected by Hurricane Matthew, a

category 5 Atlantic hurricane that skirted the southeast US coast in October, 2016. Port

operations leading up to Hurricane Matthew and observed losses in system functionality during

and following the storm are used to quantify the impact on six case-study ports using time

dependent performance analysis. Modeling and simulation involve first modeling the port

systems on a VISSIM software platform and establishing baselines for normal port operations on

the waterside and landside in the southeast region, based on archived statistics. Impacts of

Hurricane Matthew on the six case-study ports (Port of Miami, FL; Port Everglades, FL; Port of

Jacksonville, FL; Port Canaveral, FL; Port of Savannah, GA; Port of Charleston, SC) are

considered. Baseline operations are established in terms of port service and throughput at these

ports. Simulations of baseline vessel operations are conducted using the Monte Carlo simulation

approach of Inverse Transform Sampling. Once the baseline models are established, disruption

due to Hurricane Matthew is simulated to determine the consequences of the disruption. The

predictions are compared with available observed data for the event. The tool has the following

primary benefits: 1) Improvements made using the tool for hurricane preparedness, planning and

COOP at ports will lead to reduction in vulnerability of ports to hurricane events; 2) Prediction

of consequences of disruption at regional ports due to a hurricane event, validated with actual

impact of Hurricane Matthew, will lead to reduction in uncertainty of the impact of such an

event; and 3) Capability to predict and compare consequences resulting from various responses

to a hurricane event will lead to better decision-making in responding to a hurricane event.

A kick-off meeting with the U.S. Coast Guard (USCG) was held on 7/19/2018 to determine the

project’s scope, deliverables, and the case study ports. The options for tool transition were

discussed and housing the tool in USDOT’s Freight Mobility Research Institute at FAU where it

would be made available to stakeholders was chosen as the preferred option. Additional

stakeholder engagement included a well-attended stakeholder workshop at FAU on March 29,

2019, periodic conference call discussions with USCG personnel and visits to ports. The

engagement with the USCG, Department of Homeland Security (DHS) Centers of Excellence

(COE), and Port Authorities helped in the identification of the best-fitted data for the project.

Additionally, data on ship and ground vehicle movements were obtained from the DOT and on

vessel movement from Marine Traffic Online. The acquired data have been used to model the six

case-study ports and their baseline operations. The six ports were modeled, incorporating

consideration of ship movements during normal port conditions. All baseline data for a 30-day

window encompassing Hurricane-Matthew event were incorporated into the models and the

VISSIM platform, A hybrid model that connects the waterside and landside activities has been

established. Calibration and validation of the model has been carried out. There are 5 sets of

baseline data for every port, totaling 30 simulations in order to calibrate and validate the models.

Once the models were calibrated and validated, disruptive case scenario, based on Hurricane

5

Matthew was analyzed and compared with actual observations during the event. Two scenarios

were considered, one with the ports operating at normal capacity to clear the backlogs following

port closure, and the other with the ports operating with enhanced capacity. The impact on

containerized vessels in the region was quantified in terms of lost vessel hours and associated

financial impact.

Disclaimer Note: This material is based upon work funded by the U.S. Department of

Homeland Security under Cooperative Agreement No. 2014-ST-061-ML0001. The views and

conclusions contained in this document are those of the authors and should not be interpreted as

necessarily representing the official policies, either expressed or implied, of the U.S. Department

of Homeland Security.

6

ABBREVIATIONS AND ACRONYMS

AADT Average Annual Daily Traffic

BRIC Baseline Resilience Indicator for Communities

CDF Cumulative Distribution Function

COE Center of Excellence

CREATE Center for Risk and Economic Analysis of Terrorism Events COE

DHS Department of Homeland Security

DOT Departments of Transportation

FAU Florida Atlantic University

FECR Florida East Coast Railway

FY Fiscal Year

GA Georgia

JAXPORT Port of Jacksonville

JOC Journal of Commerce

ICTF Intermodal Container Transfer Facility

LLC Limited Liability Company

LNG Liquefied Natural Gas

MARS Methodology for Assessing Resilience for Seaports

MIT Massachusetts Institute of Technology

MSC Mediterranean Shipping Company

NAIS Nationwide Automatic Identification System

NOAA National Oceanic and Atmospheric Administration

NSF National Science Foundation

PANYNJ Port Authority of New York and New Jersey

PDF Probability Distribution Function

PIERS Port Import Export Reporting Service

PortSec Port Security Risk and Resource Management System

RBC Ring Barrier Controller

RIPS Resilient Interdependent Infrastructure Processes and Systems

RoRo Roll-on / Roll-off

SC South Carolina

TEU Twenty-foot Equivalent Unit

UK United Kingdom

US United States

7

1.0 INTRODUCTION

Ports are a vital part of the infrastructure for many nations. In 2014, seaports contributed

to 26 percent of the United States’ $17.4 trillion economy. Ports help to deliver essential goods

including food and gasoline to major distribution hubs to be sent throughout the country. US

Ports employ 23.1 million people and contributes $1.1 trillion dollars to personal wages and local

consumption (American Association of Port Authorities, 2015). In the United States there are 29

ports on the West Coast and 16 between the East Coast and the Gulf of Mexico (Welshans,

2015). Ten metropolitan ports across the country account for 60 percent of international goods

arriving in the country by sea, air, and road (Tomer & Kane, 2015).

US ports and container/intermodal terminals are critical links in the marine transportation

system. Disruption at a port can have a crippling economic effect in the coastal zone as well as

the rest of the nation. Ports are vulnerable to natural disasters since they are fixed, publicly

accessible entities. Port stakeholders have a vested interest in the long-term function and viability

of ports, but no standardized measures for performance or resilience exist for regional ports. In

the last 26 years, sea levels have risen 2.6 inches (NOAA, 2008). With rising sea levels, major

hurricanes (category three or higher) in the Atlantic have increased 74 percent (NOAA, 2015).

The increase in major storms has made the need for resilient marine transportation systems even

more vital. The proximity of ports to other major bodies is affected by storm surge and changing

currents and tides. Since ports are such a vital part of the economy, they are potential targets for

those seeking to do harm. In addition to the economic impact that a port disruption would cause,

the environmental effects that could occur within the waters would also threaten the local

ecology. Ports that service interconnected channels can also form queues and hinder access to

upstream facilities. This can have a major impact on capacity and the overall level of service

provided by the port system.

Hurricanes, oil spills, and labor disputes can all be sources of port disruptions. Hurricane

Sandy in October 2012 closed the Port of New York/New Jersey for over a week from full

operation. The hurricane caused flooding, loss of power, and damages to the port that prevented

the ports from reopening immediately. It was estimated by the Port Authority of New York and

New Jersey (PANYNJ) that the port closure cost $170 million (Smythe, 2013). Between the time

the port partially reopened (three days after landfall) and the time the port returned to full

operation (eight days after landfall), dwell times of vessels trying to enter the port climbed as

high as 50 hours (Wolshon, Parr, Farhadi, & Mitchell, 2018).

In September 2017, Hurricane Maria closed the Port of San Juan for three days until 11

cargo ships were allowed to enter with supplies for the devastated island. Eight days after the

initial closing 10,000 shipping containers remained stranded in the port unable to navigate the

island due to the damage to the island’s infrastructure making the roadway network impassible

for truck drivers to arrive at the port and for the trucks to leave the port to deliver the supplies.

(Gillespie, Romo, & Santana, 2017).

A labor dispute at the Port of Long Beach Labor affected 29 ports along the West Coast

of the United States and cost $7 billion in 2015. Labor disputes that close ports can cripple an

economy that shipped 240 million tons of goods in 2013 alone (DePillis, 2015). As the number

and size of ships increase, the processing time of goods in the ports decreases. Containers from

8

Asia remained anchored outside the port for 10 days before they could enter and unload

(Welshans, 2015). Even after the port reopened to normal operations it took several days to

return to normal operating level of service. Traffic congestion also affects ports. In 2014 truck

drivers waited four hours to pick up a container to transport from the port (DePillis, 2015).

In March 2014, the Galveston Bay port was closed when 168,000 gallons of fuel spilled

following a collision of a tank-barge with a long bulk carrier. The fuel spill forced the port to

close for cleanup and resulted in a large queue of vessels waiting to enter the port for four days.

Dwell times during the port’s closure exceeded 120 hours (Wolshon et al., 2018). In May 2019,

another oil spill occurred in the Houston Ship Channel depositing 9,000 gallons into the

waterway. The Houston Ship Channel is a “lifeline to the Gulf of Mexico and the foreign

markets” (Cunningham, 2019). Its closure to allow for cleanup had major implications for the

port industry in the area for both incidents.

The National Science Foundation (NSF) defines resiliency as “the ability to prepare and

plan for, absorb, recover from, or more successfully adapt to actual or potential adverse events”

(Resilient Interdependent Infrastructure Processes and Systems (RIPS), 2006). Resilient

infrastructure is more than clearing debris from roadways after a storm, it is providing the means

to deliver the essential goods and supplies needed to safely and quickly recover from a storm,

attack, or other major disruption. Resilient ports will help to create a shorter disruptive period

and a faster time to recovery of the physical infrastructure and economic livelihood of the ports.

There is an evident need to better understand the relationship between port operations,

disruptions, and resiliency on both a local and regional level.

1.2 Research Goal, Objectives, and Broader Impact

An approach to measuring resilience must be adaptable to the specific needs of the

community using it, which quickly renders a national-scale resilience metric nearly impossible.

Driven by global economic forces, ports have unique needs that should inform indicators to

assess resilience over time. Quantitative methods and tools, stemming from engineering science

and vulnerability studies, provide quick assessments of “resilience” at broad spatial scales, but

do not dip below the surface into local scale, place-based, community resilience. Qualitative

methods, on the other hand, help answer research questions that cannot be addressed with

numerical data and dive into questions of attitude, perception, and social interaction.

The goal of this research is to demonstrate the utility of a predictive port resiliency

assessment tool. The developed tool involves VISSIM based hybrid multimodal simulation that

analyzes port operations and provides a quantifiable assessment of local and regional resiliency.

VISSIM is a state-of-the-art microscopic transportation simulation model commercially

developed by VISION. The application of this tool is shown on six ports in the Southeast US.

The waterside port simulation models have been developed using vessel Automatic Identification

System (AIS) data and programed within a VISSIM simulation of landside operations. This

hybrid modeling approach is used to visualize vessels and allows them to interact in both time

and space with each other and landside port infrastructure. VISSIM also provides a means of

analyzing vessel queuing and data extraction. Local and regional resiliency is quantified through

the analysis of time-dependent resiliency plots and used as a performance measure in this study.

The utility of the predictive port resiliency assessment is demonstrated in response to Hurricane

9

Matthew. However, the novel procedure described here can be applied to any local or regional

port hazard.

This research has expanded the understanding of port resiliency and enhanced the

knowledge in the development of a stakeholder-focused tool to improve regional resilience. This

research used knowledge, innovation, and education, as well as modeling and simulation, in

support of assessment of port resiliency. Given the nature of resilience as a dynamic process, this

study considered strategies for managing identified risk. The approach used allows assessment of

port resilience and improvement in understanding of the consequences of hurricane events at the

ports and the intermodal facilities in a region, in support of seeking improvements in the regional

preparedness of interconnected port systems. The application of the predicative port resiliency

assessment tool can lead to more informed decision making for emergency managers and port

officials. Ideally, the developed tool will be used to evaluate objectively the resiliency of local

and regional ports in response to policy, protective actions decisions, and infrastructure

improvements. The dynamic simulation architecture allows for the evaluation of various

hazards, timescales, and regions. Port planners may also find this tool useful to demonstrate the

benefits of investment on resiliency using a quantifiable metrics.

The end-user stakeholders were engaged early to develop requirements for the tool,

gather available data and identify points-of-contract for obtaining feedback during the course of

the project. The data was obtained from state departments of transportation and online data

procurement resources. In conjunction with port authorities, US Coast Guard, DHS Centers of

Excellence (COE) and other stakeholders, the scope and the requirements for the methodology

was defined. Requirements to successfully quantify resiliency of the port system from a major

hurricane event and rate of recovery from it will include the following:

(i) Assessment of the level of threat

(ii) Assessment of vulnerabilities of the port system

(iii) Determination of standard operating procedure in response to a disruptive event

The completed tool has the following attributes:

(i) Provide means to reduce risk

(ii) Identify, assess, and monitor disaster risks and improve early warning systems

(iii) Build a culture of safety and resiliency at all levels through the use of knowledge,

innovation, and education

(iv) Reduce consequences from underlying risk factors

(v) Improve disaster preparedness of ports and its water and landside capacity distribution

(vi) Speed the post-disaster recover

(vii) Facilitate coordination of resumption of commercial service and relief activities

(viii) Improve interagency coordination and communication

The following section of this report highlights the relevant prior literature in the areas of port

resiliency and modeling. This is followed by a description of stakeholder engagement and a

description of the six ports used as part of the regional case study. The modeling and resiliency

10

assessment methodology is then presented followed by the analysis, results, and discussion and

conclusions.

2.0 LITERATURE REVIEW

A number of research efforts underway have focused on the resiliency of coastal

infrastructure as well as the transportation network. These include DHS funded work at

the Critical Infrastructure Resilience Institute (CIRI), CREATE COE, and the Coastal Resilience

Center (University of North Carolina-Chapel Hill), decision-support tools for flood risks at

RAND Institute; as well as lessons learned from previous storm events such as Hurricane Sandy.

This tool builds on and is complementary to these research efforts.

.

2.1 US Coast Guard Instructions

The United States Coast Guard (USCG) has well defined Pre and Post-Storm instructions

for a major weather event on the Eastern Seaboard of the country, they are outlined below. The

Coast Guard’s top priority is the safety of life during all recovery operations.

The USCG Pre-Storm instruction consists of 4 different hurricane conditions (Whiskey,

X-Ray, Yankee, and Zulu). First, Whiskey consists of the first 72 hours of the weather event,

where the Coast Guard conducts port surveys and implements containers stacking protocols.

Second, X-Ray is within 48 hours, where all ocean-going vessels over 500 GT are required to

depart port and a “remaining in-port plan” must be submitted to the Coast Guard if the vessels

are unable to depart. The Yankee condition takes place within 24 hours of the weather event,

which closes the port to all inbound traffic, all-cargo operations. Additionally, drawbridges may

cease operation within 8 hours of the predicted tropical winds. The last condition is Zulu which

occurs 12 hours prior and throughout the duration of the weather event. This last condition

suspends all port waterfront operations and both (port and waterfront) facilities remain closed

until the passage of the storm.

The USCG post-storm plan involves meeting and verifying several conditions prior to the

port opening. The following five conditions must be met, at the bare minimum, prior to the port

opening. The first condition involves conducting a damage assessment to ports and waterways

completed in conjunction with the Army Corps. The second condition is the verification of the

correct location of the aids to navigation. The third condition is checking the status and

conditions of all drawbridges. The fourth condition is the re-establishment of required port

security measures in accordance with respective port security plans; it involves Customs and

Border Protection to be ready to process all cargo and personnel. The fifth condition requires the

evaluation of oil spills or hazardous material release in the port area and to identify the potential

sources.

Depending on the severity of the weather event, the Captain of the Port may issue

Captain of the Port Orders called a “Soft” Port Opening. This soft port opening may consist of

the following restrictions depending on the damage to the port. They may establish safety zones

or restrict transits to “daylight only” and additionally may require additional tug escorts.

2.2 Resiliency Assessment

11

Adam Rose and Shy-Yi Liao (2005) address regional resilience towards disasters in their paper

“Modeling Regional Economic Resilience to Disasters: A Computable General Equilibrium

Analysis of Water Service Disruptions”. They utilize the computable general equilibrium

modeling approach in order to estimate the regional economic impacts of earthquakes and other

disasters that can cause supply chain disruptions. Some of the main functions of this model

included operational definitions of individual and regional resilience, identification of production

function parameters and development of the algorithms for recalibrating production functions to

data.

In addition, Mo Mansouri, Roshanak Nilchiani and Alsi Mostashari (2011) conducted a research

project titled “A Policy Making Framework for Resilient Port Infrastructure Systems”. In their

work they developed a Risk Management-based Decision Analysis framework, with the goal of

forming a systematic process for making strategic and investment decisions in case of

disruptions. The disruption cases considered ranged from natural disasters, to organizational,

technological and human factors. Their approach can help to identify common elements of

uncertainty in port systems, evaluate the costs incurred with various potential failures and with

investing in resilience strategies.

In their article titled, “Resilience Framework for Ports and Other Intermodal Components,”

Rahul Nair, Hakob Avetisyan and Elise Miller-Hooks (2011) discuss ports and intermodal

freight systems, highlighting the dangers that hinder cargo transportation and the infrastructure’s

vulnerability to disasters. They quantify resilience as the post disruption fraction of demand that

can be satisfied while using specific available resources and managing to maintain a prescribed

level of service. Additionally, they employ their concept on a system level and propose a generic

framework for its application in intermodal facilities. In another article, Elise Miller-Hooks,

along with Xiaodong Zhang and Reza Faturechi (Miller-Hooks and Faturechi, 2012) conducted

another research study related to resilience named “Measuring and Maximizing Resilience of

Freight Transportation Networks”. Their model, apart from measuring resilience levels of a

freight network, includes optimal setting of actions and allocation of budget between

preparedness and recovery activities under level-of-service constraints.

The National Center for Risk and Economic Analysis of Terrorism Events at the University of

Southern California spearheaded a project titled “PortSec: Port Security Risk and Resource

Management System” (Orosz, 2011). Its objective was to create a system for risk assessment and

security resource allocation for various dangers that hinder seaport operations. This decision

matrix is used by the port authorities to manage and balance the increasing safety restrictions.

This matrix allows the Port Authorities to maximize business throughput and minimize

environmental impacts. The project has two main uses, strategic and tactical. Strategic usage

includes the creation of tools for evaluation of the cost-benefit by adding/modifying new port

counter-measures. On the other hand, the tactical usage of the tool provides up-to-date risk

assessment for both identified areas of interest and for the overall port complex.

The Center for Transportation & Logistics at the Massachusetts Institute of Technology (MIT)

conducted a multi-year Port Resilience project (Rice and Trepte, 2013). The goal of the study

was to estimate the capacity required to absorb various failures of United States ports. The

project included a port capacity analysis, port failure mode analysis and a detailed port resilience

survey. In addition, they developed a platform called MIT Port Mapper, which is designed to

identify U.S. ports that can potentially absorb cargo in the event of a port disruption. The user

12

chooses the state he wishes to examine and either all or a portion of the state’s ports. In addition,

the platform is used for gathering information on the type of materials handled in each port (e.g.

radioactive, containers); the data in the platform was obtained from the Army Corps of

Engineers.

Moreover, the Americas Relief Team (2013) in collaboration with FedEx conducted a project

titled “Port Resiliency Program” Its objective was the preparation of airports and seaports in the

Caribbean and Latin America to be more resilient in the face of natural disasters by applying

lessons learned in Hurricane Katrina and the Haiti earthquake. Their approach for achieving their

goal comprised of three main steps: Initial self-assessment by the airport or seaport; planning of

a workshop in Miami to identify gaps and training needs in sea and air port operations; and a site

visit to present targeted training and a tabletop exercise to assess the preparedness of the airport

or seaport.

A study conducted by Tiffany C. Smythe (2013), from the Center for Maritime Policy and

Strategy of the U.S. Coast Guard Academy, titled “Assessing the Impacts of Hurricane Sandy on

the Port of New York and New Jersey’s Maritime Responders and Response Infrastructure,”

focused on Hurricane Sandy. The goal of the study was to identify “lessons learned” from this

hurricane, in order to educate the maritime community in the necessary actions required to

mitigate impact during future storm events. The methodology used for achieving the goal

included three main steps: meetings with the U.S. Coast Guard, data collection via participation

in meetings and semi-structured interviews with key informants, and qualitative data analysis of

the interview content. In addition, the research aimed to lay the groundwork for larger-scale and

longer-term studies related to coastal storms and port resilience planning.

Other qualitative analyses of port disruptions due to hurricanes have been undertaken to study

stakeholder perceptions (Becker, Matson, Fischer, & Mastrandrea, 2015). This article proposed a

storm impact typology for two ports (Gulfport, Mississippi and Providence, Rhode Island) to

include direct damages, indirect costs, and intangible consequences. The authors found that

formal planning did not address many stakeholder concerns, particularly the impacts of

intangible consequences that are borne by a large number of stakeholders and society at large.

In addition, an important area of research in the field of port resilience deals with the

identification of the costs associated with port disruptions. Adam Rose and Dan Wei (2013), in

their paper called “Estimating the Economic Consequences of a Port Shutdown: The Special

Role of Resilience” address this matter. They developed a demand and supply-driven

methodology that takes into account imports, exports and the major types of resilience in terms

of alternative options. The study was successful in developing a tool to estimate the total

economic consequences of a seaport disruption. After applying their approach to a 90-day

disruption at the seaports of Beaumont and Port Arthur, Texas, they concluded that a carefully

thought out resilience plan can reduce the impacts of disruption by as much as 70%.

Studies regarding port resilience have also been conducted outside the U.S. An example is a

paper from Andrew Grainger and Kamal Achuthan (2014) from the University of Nottingham, as

part of a collaborative project with the United Kingdom Department of Transport, titled “Port

Resilience: a Primer”. The study focused on the importance of U.K. ports, various vulnerability

13

issues often encountered in them, preparedness methods followed to address those problems and

various actions that can be taken in order to improve port resilience. Some of these actions

include the development and adoption of strict planning and business continuity standards,

development of simulation tools that can help understand and predict specific events taking place

and identification of incentive mechanisms to ensure stakeholder interests towards resilience.

Hui Shan Loh and Vinh Van Thai (2014) in their paper “Managing Port-Related Supply Chain

Disruptions: A Conceptual Paper” focused on the management side of port resiliency. They

developed a management model that addresses a full set of operational risks from a holistic

perspective, and in connection with various supply chain disruptions. Therefore, their model

identifies the necessary actions taken from ports in order to minimize port-related supply chain

disruptions. The proposed approach incorporates the theories of risk, quality and business

continuity management in order to make decisions in the institutional bearings, management

policies and operations actions related to port disruptions.

Moreover, a journal paper that deals with the quantification of resilience was authored by

Raghav Pant, Kash Barker, Jose Emmanuel Ramirez-Marque, Claudio M. Rocco (2014) and

titled “Stochastic measures of resilience and their application to container terminals”. In their

work they modeled the system resilience as a function of both vulnerability and recoverability,

while also incorporating aspects of stochasticity and uncertainty in terms such as time to total

system restoration and time to full system service resilience. The resiliency decision making

framework created includes commodity flows at a port, full or partial terminal closures due to

disruptive events and restoration activities and was applied in a case study at the port of Catoosa

in Oklahoma.

The National Cooperative Freight Research Program of the Transportation Research Board

developed a large project called “Making U.S. Ports Resilient as Part of Extended Intermodal

Supply Chains” (Southworth, Hayes, McLeod, & Strauss-Wieder, 2014). The main project tasks

first included a thorough literature review on past disruption events that affected port operations,

emphasizing the actions taken to tackle the problem and limit the extent of the disruption. Next,

interviews with port operations, truck, rail, and ocean vessel carriers were conducted in order to

understand their opinions on current levels of port resiliency, as well as on what are the best

means of enhancing resiliency and speeding recovery should a disruption occur. Later, two

detailed case studies of port disruptions were developed, the impacts of the superstorm Sandy’s

on the major East Coast ports and the extended lock closures along the Columbia River System

in the Pacific Northwest. Last, the team developed guidelines suitable for public-sector decision

makers who might become involved in a disruption recovery event.

An interesting project was conducted by the Stevens Institute of Technology, and funded from

University Transportation Research Center, Region II, with the title “Port Resilience:

Overcoming Threats to Maritime Infrastructure and Operations from Climate Change”. The

study states that the growing concerns about climate change and severe weather events occurring

has transformed the area of port and coastal resilience into an important component in operations

planning. The principal objective of the project is the creation of a standardized framework for

the improvement of resilience in ports and transportation systems, via the integration of physical

infrastructure and social systems. Stakeholder interviews, and workshops were organized that

14

improved social awareness and identified the most important problems encountered while

dealing with disruptions. Some of the solutions identified was the implementation of strict design

standards, the organization of the transport systems as a whole, in terms of the entire supply

chain, and to look beyond local operations. Last, the project suggested a coordinated

organizational scheme at the state and regional level that can assist in the interaction towards the

landside operations and water side logistics teams throughout the whole disruption cycle.

Another project with significant value to the problem addressed in our study was conducted from

the Centre for Transport Studies of the University College London (Achuthan et al., 2015), titled

“Resilience of the Food Supply to Port Flooding on East Coast”. The study argues that since the

UK imports more than 40% of its food and drink supplies, with most of it arriving by sea, it is of

utmost importance for port systems to be able and adopt good resilience plans. The methodology

followed in the project consisted of engagement with stakeholders, modelling and analysis of the

UK ports and shipping import functions, and the development of disruption scenarios using

simulation. Some of the key findings extracted from the project were the degree of the disruption

at a port would vary according to the food type moved and the shipment method, and that

rerouting RoRo and container vessels to other available ports can potentially reduce the impacts

of the event to almost half.

Furthermore, the Gulf of Mexico Alliance (Morris, 2016) conducted a study named “Ports

Resilience Index: Three Case Studies in the Gulf of Mexico”. The objective of the project is the

production of a simple and easily implementable regional tool that port and marine transportation

authorities can use to evaluate and assess their level of resilience, as well as predict their ability

to achieve an acceptable level of service during and after major weather events. The Ports

Resilience Index is constructed using the Delphi Method, commonly used for quantifying

variables of uncertainty and reaching a statistical consensus. The case studies considered were

the Port of Corpus Christi in Texas, the Port of Pascagoula in Mississippi and the Port of Lake

Charles in Louisiana.

A research paper from Justice et al.(2016), named “US Container Port Resilience in a Complex

and Dynamic World” addressed the problem of how container ports in the U.S. can potentially

be affected by various negative events and how they can implement resilience practices to

counteract the issues. The authors emphasized the importance of resilience as a way of dealing

with uncertainties, while also mentioning that due to these potential changes, the ‘business as

usual’ approach adopted by most organizations may not be able to guarantee successful port

operations. Last, they state that in order to manage and encompass resilience, innovative and

creative methods are required.

Hong Chen, Kevin Cullinane and Nan Liu (2017) also dealt with the subject of measuring

resilience in transportation networks, in their paper titled “Developing a Model for Measuring

the Resilience of a Port-Hinterland Container Transportation Network”. In their study, after

developing their own definition of resilience in transportation networks and port operations, they

developed a model to quantify resilience while incorporating links, nodes, cost, time and port

capacity from the perspective of shippers. They considered a single seaport in their model and

applied their methodology to the Gothenburg port and part of its hinterland.

15

Hyungmin Cho and Heekyung Park (2017), co-authored a paper titled “Constructing Resilience

Model of Port Infrastructure Based on System Dynamics”, creating another study that focused on

building resilience models of ports. In their work, they state that since port infrastructure and

operations are complex processes and difficult to analyze all their components, a systemic

approach can prove efficient. Their system dynamics model incorporates the cargo process as its

performance level and identifies the elements corresponding to various resilience attributes.

Additionally, the model includes factors such as changes in cargo volumes and financial states,

and with the application of different disruption scenarios, is used as a method of comparing the

resilience levels of port infrastructure.

A relevant work that focused on the impacts of hurricanes in port operations was conducted by

Touzinsky et al. (2018), titled “Using Empirical Data to Quantify Port Resilience: Hurricane

Matthew and the Southeastern Seaboard”. In their study they used Automatic Identification

System based vessel arrival data on three case study ports hit by hurricane Matthew, Charleston,

Savannah and Jacksonville. Their goal was to calculate cumulative dwell times and net vessel

counts in order to simulate and quantify the behavior of the system during all the main stages of

the hurricane. These stages included pre-storm, preparedness, resistance, recovery and post-

storm. In each stage, they used Bayesian analysis for understanding the system performance

variations over the whole hurricane incident time.

2.3 Port Modeling

Port modelling takes on multiple aspects to obtain the most realistic model of the area. Since

ports are a transfer point for cargo, they have unique characteristics and can be affected by both

the landside and seaside of the operation by disruptions. Ports are a large part of the economy

requiring them to be resilient in the face of disruption to continue to provide for the country. The

National Science Foundation (NSF) defines resiliency as “the ability to prepare and plan for,

absorb, recover from, or more successfully adapt to actual or potential adverse events” (Resilient

Interdependent Infrastructure Processes and Systems (RIPS), 2006).

According to John, Yang, Riahi, and Wang, “when critical maritime systems do not have the

robustness to recover in the face of disruption, they present themselves as attractive targets to

terrorism-related attacks” and “disruptions at any point within their operation could potentially

result in catastrophic and disastrous consequences” (John et al., 2015). Alyami, Yang, Riahi,

Bonsall, and Wang agreed that resiliency is important because “ safe and reliable operations are

of great significance for the protection of human life and health, the environment, and the

economy” and early detection of hazards is crucial in avoiding performance degradation and

damage to human life and property” (Alyami et al., 2016).

Mangan, Lalwani, and Gardner (2002) point out that reliability has been highly ranked in port

choice by many researchers. Murphy and Hall ranked it as the number one most important

variable in predictive port modeling in 1995 and Cullinane and Toy ranked it as number three in

2000 (Mangan, Lalwani, and Gardner, 2002). While reliability and resiliency are conceptually

different, they can result in similar outcomes. Ports must be reliable in their ability to operate

prior to, during, and following adverse conditions; therefore, being resilient.

16

Magala and Sammons state the “in a highly competitive and rapidly globalizing economy of

today, the integration of supply chains is taking pace and ports are increasingly competing not as

individual firms but rather as firms within supply chains” (Magala and Sammons, 2008). This is

an important aspect of port modelling because when modelling a port just the movements alone

within the port cannot be solely examined, but the interaction with the vessels offshore and the

rail or trucks used to move the cargo from the port to its destination. Magala and Sammons

reasoned that by modeling them as “an integrated supply chain the opportunity to reduce

vulnerability to competition by providing the port with complementary resources and capabilities

needed to compete more effectively in the marketplace” (Magala and Sammons, 2008).

According to Al-Deek, “seaports are primary generators of freight traffic, both truck and rail (Al-

Deek, 2001). This supports ideas stated by Magala and Sammons that when modelling a port the

entire system must be examined. Al-Deek states that the trip generation and modal split aspects

of the “four-step transportation planning process” must still be done to relate the number of truck

and rail trips generated by the vessels (Al-Deek, 2001). Traffic movements within a port can be

as high as 2000 per day and the number is expected to continue to increase (Debnath, Chin,

Haque, 2011). With such a high number of traffic movements “terminals must have the facilities

to provide adequate level of service” (Kozan, 1996).

Le Tixerant, Le Guyader, Gourmelon, and Queffelec note that this in a “multiplying,

diversifying, and intensifying sea-related use” and with that an “increase in traffic linked to

maritime shipping” and other activities (Le Tixerant et al., 2018). Due to the ever-changing

nature of ports the variability in delays must be examined when modeling ports since level of

service is dependent on it (Kozan, 1996). AIS (Automated Identification Systems) can be used to

help model ports by allowing for classifying vessel patterns (Chen, Xue, Wu, Qin, Liu, & Chen,

2018).

Kozan and Preston acknowledge that the arrival time of a vessel to the port, also known as the

berthing time, “accounts for a considerable portion of the journey;” therefore, it is an important

aspect to be included in port modelling (Kozan and Preston, 2006). Due to the complexity of

ports and “the dynamic nature of the environment, a large number of timely decisions have to be

continuously reviewed in accordance with the changing conditions of the system” (Kozan and

Preston, 2006). As vessels enter a port they are received until full capacity and with minimum

queuing time that is determined using the average waiting time, optimum number of berths, and

the maximum numbers of ship waiting to enter (Costea, Ticu, & Mishkoy, 2019).

Howard, Bragen, Burke, and Love discuss the need for a port to be able to “deploy an armed

cavalry division through the port, given a specified allocation of port resources (Howard, Bragen,

Burke, and Love, 2004). Howard et al. used PORTSIM 5 to model this and to help “identify

ports for which additional infrastructure (newly constructed or lease from the port) can speed the

process of deployment” (Howard et. al, 2004). This is important when examining the resiliency

of ports and what is available prior to and following an event.

Vessel Traffic Service (VTS) uses vessel movements to predict movement on waterways through

“collection, verification, organization and dissemination of information” (U.S. Coast Guard,

2018). VTS is important in port modeling because it helps to identify vessels that have

17

navigation limitations that may need assistance to enter the port. Vessels needing assistance

affect the operations of ports greatly. Vessel navigation behavior determined by VTS is

compared to existing port models that allow for demand to be examined within the port (Olba,

Daamen, Vellinga, & Hoogendoorn, 2018). VTS allows for more accurate port modeling by

helping to determine if vessels abide by the International Maritime Organization (IMO)

movement rules within the ports or if the follow different and often dangerous paths that would

increase risk and decrease safety (Olba et al., 2018).

3.0 STAKEHOLDER ENGAGEMENT AND CASE STUDY PORTS

US Coast Guard, DHS COEs, and Port Authorities are the primary stakeholders when

responding to natural disasters and port operations. These stakeholders were engaged with

throughout the research process and helped in the identification of the study areas, data

requirements, project scope, and deliverables. A series of interviews, meetings with stakeholders,

and workshops were held with stakeholders to guide the project beginning with a kick-off

meeting and combination in the submission of the final report.

Meetings with officials from DHS and the US Coast Guard were held in order to determine the

scope of the project and its benefits for the stakeholder purposes. These meetings helped form a

better view of the problem and the needs of the stakeholders so that the results of the research

will be useful in making port operations more efficient during an emergency event. Further

information about the operations of the case studies was collected through meetings with the Port

Managers of each port and visits to ports.

On March 29, 2019, major stakeholders with interest in infrastructure resilience and impacts

convened for our Supply Chain Sustainability and Transportation Resilience workshop, in Dania

Beach, Florida. This workshop focused on both supply chain sustainability and transportation

resilience, with a wide range of presentations, that spun from framework proposals to

transportation and operations management-based research. Acknowledging the breadth of issues

faced by ports and the communities they affect. The port stakeholders’ workshop in resilience

built on outcomes for two theme-based listening sessions and collaboratively exploreed

challenges and opportunities facing ports and neighboring communities. The key goal from the

workshop was to promote engagement with other professionals, experts, and stakeholders to

share expertise, ideas and actions that address the many opportunities and challenges faced by

our nation’s major transportation infrastructure, intermodal facilities/ports and communities.

3.1 Case-Study Ports

The project considered six different ports that were hit by Hurricane Matthew in 2016, (Port of

Miami, FL; Port Everglades, FL; Port of Jacksonville, FL; Port of Savannah, GA; Charleston,

SC; Port Canaveral, FL). At these ports, pertinent data were collected during Hurricane Matthew.

Hurricane Matthew was unique in the fact that it had a direct impact on all six ports. In other

words, all six ports were closed at some point during this weather event. Once the models were

built, the data from Hurricane Matthew were used to validate the output of these models. This

validated model was then used to assess the regional resiliency of these six ports. Each of these

ports are vital to the Southeastern United States’ economy and their resiliency is paramount.

18

They all individually contribute, although in different ways, to this transportation network.

Outlined below are the specifics of each port, along with their contributions to the economy of

the Southeastern United States.

3.1.1 Port of Miami “Port of Miami is one of the world’s leading hubs for global commerce. Its

gateway location in the center of the Western Hemisphere makes the Port a significant conduit

for international trade. Port of Miami stands as the U.S. container port closest to the Panama

Canal and is the only major logistics hub south of Virginia capable of handling fully laden post-

Panamax vessels, providing shippers fast access to Florida’s booming local consumer base and

the entire U.S. market. With seven container ports and nine cruise terminals, Port of Miami plays

a major role in the international shipping industry.”

“Latin America and the Caribbean makeup 49% of Port of Miami’s trade region in 2018. With

the completion of the Deep Dredge Project, trade with Asia will increase from its current 33% as

Port of Miami benefits from a shift in trade from West Coast to East Coast ports. As a staple in

the global economy, Port of Miami’s remaining trade region comprises of the Europe, Middle

East, India, and Africa. In 2018, the total TEU capacity estimated to 1.1 million twenty-foot

equivalent units of containerized cargo (TEUs), a 37% increase from 2014. The total value of the

economic impact created by cargo containers moving via Port of Miami is estimated at $35

billion dollars to the State of Florida. This economic impact includes increased value added

during the production of export cargo, as well as transportation, warehousing, and retail

distribution activities for import cargo. This import and export activity generates state income

equaling $10 billion dollars, and state and local taxes at approximately $2 billion dollars

annually. Equipped with 13 ship-to-shore cranes including six super Post-Panamax, the largest in

the Southeastern U.S. and on dock intermodal freight rail connects the port with the national rail

system moving goods throughout Florida and the continental U.S. reaching 70% of America’s

market within four days.”

“Known as the “Cruise Capital of the World”, Port of Miami is homeport to an unprecedented 22

Cruise lines and 55 of the industry’s most innovative ships. Port of Miami's cruise terminals --

among the most modern in the world -- have been designed to quickly move passengers from

land to sea. Port of Miami contributes approximately $43 billion and more than 334,500 jobs to

Florida’s economy.”

“In October 2018, Royal Caribbean International officially opened Cruise Terminal A, the

Crown of Miami. The new facility serves as homeport to the world’s largest cruise ship, Royal’s

Symphony of the Seas. April 2018, Port of Miami and Norwegian Cruise Line’s celebrated the

groundbreaking for the new cruise terminal Terminal B, The Pearl of Miami. The facility will

accommodate vessels carrying up to 5,000 cruise passengers. It is scheduled to open February

2020. MSC Cruises’ new facilities are in the works, and an MOU with Disney Cruise Line to

expand with two cruise ships and a possible new cruise terminal was approved. Virgin Voyages

has also proposed building a new terminal in 2021 for its new cruise ship, the Scarlet Lady

Source: https://www.miamidade.gov/portmiami

3.1.2 Port Everglades “Known as an “economic powerhouse, Port Everglades is one of the

busiest cruise ports in the world. It is a leading container port in Florida and among the most

19

active cargo ports in the United State and is South Florida’s main seaport for receiving petroleum

products including gasoline and jet fuel.”

“With more than 6.6 million tons of containerized cargo in Fiscal Year 2017, Port Everglades is

the 10th busiest container port in the nation, according to the Port Import Export Reporting

Service (PIERS) published by JOC Group Inc. Port of choice for more than 20 shipping lines,

Port Everglades is ideal for moving products such as fruit, vegetables, automobiles and apparel

to and from Central America, the Caribbean, South America, Europe and even the Far East.

Poised with superior customer service and state-of-the-art equipment, Port Everglades has

increased its capacity to handle containerized cargo from 880 thousand TEU’s to approximately

1.1 million.”

“Equipped with 9 cranes capable of handling a full range of cargo, Port Everglades has kept pace

with Florida’s construction and population demand by moving various bulk/break bulk material

to include: imported and exported cement, lumber, steel rebar, cement, aggregate, tallow, and

gypsum, handling 1.5 million units.”

“The 43-acre near-dock Intermodal Container Transfer Facility (ICTF) at Port Everglades is

operated by Florida East Coast Railway (FECR). The ICTF's advantageous location adjacent to

the Port's docks allows international containers to be quickly transferred between ship and rail,

reducing congestion on interstate highways and local roadways and reducing harmful air

emissions by diverting an estimated 180,000 trucks from the roads by the year 2029.”

“With annual passenger counts of nearly 4 million, Port Everglades holds the distinction of being

among the three busiest cruise ports in the world with 10 cruise lines and one daily ferry

operator. The total number of cruise and ferry passengers increased by 1 percent for a total of

3,863,662 passengers in FY2017. Multi-day cruise passenger numbers increased by 2 percent

from 3,680,549 in FY2016 to 3,738,252 passengers in FY2017. Daily passenger numbers

dropped slightly from 145,866 to 125,410 in FY2017.”

Source: http://www.porteverglades.net

3.1.3 Port of Jacksonville (JAXPORT): “JAXPORT offers service from dozens of ocean

carriers, and over the past few years, JAXPORT has added new direct port calls and more than

100 transshipment port calls worldwide. With 1,600+ reefer plugs on dock, and dedicated cold

chain supply chain services throughout the region, the Port of Jacksonville is uniquely positioned

to safely handle temperature-controlled freight. It offers multiple berths, skilled labor and three

major auto processors to move your cars, trucks, SUVs, motorcycles and recreational boats on

time and in dealer-ready condition. JAXPORT is equipped with ten non-containerized berths,

with 1 million square feet of on-dock warehousing is advantageous to moving breakbulk cargo

like lumber, rolls of paper, wood pulp, steel and other metals. JAXPORT offers bulk shippers

more than 40 acres of dedicated dry bulk space and 324,000 barrels of capacity for liquid bulk

cargo shipments. Port partners are pioneers in the transportation of LNG in an off-pipeline,

efficient fuel supply chain from two tanker berths. With a single passenger terminal, Carnival

20

Cruise Lines’ 2,056-passenger Carnival Ecstasy offers year-round service from Jacksonville,

Florida.”

Source: https://www.jaxport.com

3.1.4 Port Canaveral “Port Canaveral is home to three seasonal and six year-round cruise ships

from Carnival Cruise Lines, Disney Cruise Line, Royal Caribbean International, and Norwegian

Cruise Line, and will welcome an estimated 104 visits by port-of-call vessels in Fiscal Year

2017. With more than 4.2 million revenue passengers last year, Port Canaveral is ranked as the

second-busiest cruise port in the world and serves as a popular homeport and port of call for

some of the largest ships afloat, including Royal Caribbean International’s Oasis of the Seas and

Norwegian Cruise Line’s Norwegian Epic. In November 2016, both of these majestic vessels

joined the Canaveral home-ported fleet.”

Port Canaveral has 12 general cargo berths, along with 1 tanker berth. Two deep-water container

and multi-purpose cargo berths make Port Canaveral the most economical and convenient ocean

gateway for containerized cargo in Florida. 1,800 feet of berthing spaces is equipped with two

ship-to-shore cranes able to handle 40.6 metric tons and move 30-40 containers per hour.

“Roll-on/Roll-off specialized shipping facilities are located at South Cargo Pier 4. This secure

paved facility covers 16 acres which includes an onsite 20,000 sq foot processing warehouse.

Port Canaveral is ideal for importing and exporting new and used RO/RO cargoes. Adjacent to

Orlando, the largest auto rental market in the world and home to several automobile and heavy

equipment auctions. AutoPort Canaveral, LLC allows for auto inland distribution to all major

Florida markets within 3 hours by truck and to over 80 million people in the Southeast US within

8 hours.”

“With unique local aerospace programs, a diverse manufacturing community and a growing

regional population, project cargoes through Port Canaveral include everything from specialized

industrial machinery to aerospace components to massive defense-related items. Two shore

cranes, in addition to a 40 metric-ton capacity mobile harbor crane are available for project

cargo. “

“Nearly four million tons of dry and liquid bulk cargo are handled annually at Port Canaveral,

including petroleum, aggregates, cement, salt, sand, and slag. Facilities feature a 2,800 foot

aggregate conveyer system with a discharge rate of 2,200 tons per hour. Special storage facilities

are available for cement, slag and petroleum.”

“In 2016, the Canaveral Port Authority completed $137 million in capital projects, including $45

million in renovations to 24-year-old Cruise Terminal 5, increasing its capacity from 2,500 to

3,500 passengers. In addition, 21-year- old Cruise Terminal 10 was renovated with a capacity

increase from 3,200 to 5,500 passengers. Cruise Terminal 8, the Disney Terminal, also received

a $2-million upgrade. In addition, the port invested more than $44 million in widening and

deepening the harbor and $15.2 million in cargo terminals and backup areas.”

“With 5.5 million tons of cargo in 2016, Port Canaveral serves as the gateway to Central Florida

and is located three hours or less from every major Florida market. Last year Port Canaveral

21

became the exclusive U.S. stop on Streamline’s Blue Stream, a weekly container service

operating from Canaveral Cargo Terminal, the first U.S. venture by Gulftainer. In addition,

Delaware-based Autoport, Inc. began handling vehicles from the port’s new auto terminal, and

Swiss aerospace company, Ruag Space USA, became the first tenant of the Port Canaveral

Logistics Center in Titusville.”

“As part of an overall expansion plan, and with a goal of accommodating larger vessels, Phase 1

of a dredging project was completed in late Fiscal Year 2016, which widened Canaveral’s 3.5-

mile channel by 100 feet and expanded the current width to 500 feet overall, and initiated the

harbor entrance deepening project.”

Sources: http://flaports.org/ports/port-canaveral/

https://www.portcanaveral.com/Cargo/Facilities

3.1.5 Port of Savannah “Port of Savannah is home to the largest single-terminal container

facility of its kind in North America, is comprised of two modern, deepwater terminals: Garden

City Terminal and Ocean Terminal. Together, these facilities exemplify the Port Authorities

exacting standards of efficiency and productivity. Garden City Terminal is the fourth busiest

container handling facilities in the United States, encompassing more than 1,200 acres and

moving millions of tons of containerized cargo annually.”

“Ocean Terminal, Savannah’s dedicated breakbulk and Roll-on / Roll-off facility, covers 200.4

acres and provides customers with more than 1.4 million square feet of covered, versatile

storage.”

“Savannah handles about 40 percent of the containerized poultry and is the second busiest

container exporter in the U.S. In FY2018, it handled 4.2 million TEUs (Twenty Foot Equivalent

Units) in throughput. FY2018 had the highest volume in the Port of Savannah’s history. The Port

of Savannah moved 8.6% of total U.S. containerized loaded cargo volume and more than 19% of

the East Coast container trade. The port handled 10% of all U.S. containerized exports in

FY2018 (USA Trade Online). The Port This port has 25 cranes, 18 berths, one container

terminals without rail access, and one with rail access. Accessible to two major interstates as

well, creating a direct four-hour drive to Atlanta, Orlando, and Charlotte that are major market

cities.”

Source: http://gaports.com/port-of-savannah

3.1.6 Port of Charleston “Port of Charleston is one of the Southeast United States' busiest

container ports. In terms of the dollar value of international shipments, the Port of Charleston

Customs district is the eight largest in the United States. In 2018, the Port of Charleston handled

almost 2.3 million TEUs of containerized cargo. In 2017, the port moved about 230k vehicles

(imports and exports) and handled about 780k tons of non-container cargo. Also the port has 19

cranes, 9 berths, 3 container terminals without rail access, and 3 with rail access. This port

generates about $53 billion dollars of South Carolina economic impact and about 187k jobs.”

22

“The major commodities handled by the Port of Charleston included consumer goods,

agricultural products, vehicles, machinery, metals, chemicals, and clay products. The major

exports leaving the Port of Charleston include paper and paperboard, wood pulp, auto parts, logs

and lumber, fabrics (including raw cotton), and general miscellaneous cargoes.”

“The Port of Charleston handles a wide range of imports. The dominant imported cargoes

entering the Port of Charleston include furniture; auto parts; sheets, blankets, and towels; fabrics;

tires and tubes; apparel and footwear; household goods; and yarns. Other imports in the Port of

Charleston include machinery parts, logs and lumber, hardware, engines, motors, and parts.”

“Over 20 marine carriers move cargo between the Port of Charleston and over 150 nations

around the world. Based on the volume of cargo, the Port of Charleston's major trading partners

include North Europe (36%), Northeast Asia (22%), India and Southeast Asia (17%), and South

America (11%). Other trading partners include the Mediterranean region, Africa, the Middle

East, Central America, and the Caribbean.”

“The Port of Charleston offers some of the deepest water in the South Atlantic region. The Port

of Charleston maintains a harbor depth of 13.7 meters (45 feet) and a channel depth of 14.3

meters (47 feet) at mean low tide, and it has a tidal lift of from 1.5 to 1.8 meters (five to six feet)

that provides from 46 to 14.6 meters (48 feet) depth for ten hours each day. The entrance channel

to the Port of Charleston is from 152 to 305 meters (500 to 1000 feet) wide.”

Sources: http://www.worldportsource.com/ports/commerce/USA_SC_Port_Charleston_248.php

http://www.scspa.com/news/sc-ports-authority-achieves-6-percent-growth-in-2018/

4.0 METHODOLOGY

Broadly, the research methodology consisted of three primary tasks. The first task was to define

an objective, quantitative approach for evaluating resiliency that would enable a robust

performance metric. The second task was developing, calibrating, and validating the hybrid

multimode simulation model of the six ports located in the Southeastern United States. When

combined with the resiliency assessment performance metric, the simulation model was used as a

predictive port resiliency tool. The final task was the development and programing of the

evaluation scenarios. The first scenario was a typical 30-day performance period to establish

base line of port operations. The second scenario was a 30-day period encompassing the

disruption of Hurricane Matthew and the recovery thereafter.

4.1 Resiliency Definition and Assessment

This research utilizes the National Science Foundation’s (NSF) definition of resilience as “the

ability to prepare and plan for, absorb, recover from, or more successfully adapt to actual or

potential adverse events”. However, while this definition is informative, it lacks a quantitative

reference. A quantification of resiliency allows systems to analyze current resiliency, track future

improvements, and model possible scenarios. Currently, there are multiple methods of

quantifying resiliency, all with their own definition for quantification. A common method used is

the scorecard method. The scorecard is a way for stakeholders to assess their resilience using

predetermined questions designed to target indicator frameworks. The Department of Homeland

23

Security has developed the Plan Integration for Resilience Scorecard method to reduce the

country’s vulnerabilities to hazards.

A similar method of resiliency quantification is the resilience matrix created by the US Army

Corps of Engineers. They utilize a 16-cell decision making matrix based on the Network Centric

Warfare doctrine of the military. This matrix produces a quantification of “poor”, “moderate” or

“good” resiliency. They also utilize a second method of quantification called the Baseline

Resilience Indicator for Communities (BRIC). This method is a composite matrix that calculates

a resiliency index value based on up to six sub-indexes. Utilizing a resilience range of “poor”,

“moderate”, “good” or “low”, “medium”, “high” is a common output for resilience

quantification. Most inputs for this method require stakeholder feedback for a variety of multiple

choice questions.

A third method of resiliency quantification utilizes stakeholder feedback to output delays and

queues in operations. This technique was created for assessing seaports by Kamal Achuthan

(2011) but has the potential for use across all transportation systems. The Methodology for

Assessing Resilience for Seaports (MARS) models the wet and dry side of port operations before

outputting delay and queue time. Stakeholders must assess the results and determine if the times

are satisfactory for their seaport.

NSF’s definition of resiliency calls for a means of measuring the system’s ability to absorb,

adapt, and recover. Figure 1 provides insight into how this can be accomplished. Let function

𝜔(𝑡) represent a direct measure of system output at any time t. System S will undergo five

distinctive states. Prior to event E (𝑡 < 𝑡𝐸), the system is operating in stable, pre-event

conditions. After event E, output decreases as the system absorbs the impact of the disruption.

Eventually, the system will stabilize as the effect of the disruption reaches its maximum impact

on functionality. While system performance is no longer decreasing, system output is still

reduced from the pre-event conditions 𝜔(𝑡𝐴) ≅ 𝜔(𝑡𝐴+1) < 𝜔(𝑡𝐸−1). The system will remain in

the disrupted state until a recovery action is taken at 𝑡 = 𝑡𝐷. The system begins to recover as

functionality is restored, 𝜔(𝑡𝐷+1) > 𝜔(𝑡𝐷). This recovery continues until the system reaches a

stable recovery at 𝑡 = 𝑡𝑅.

PRE-EVENT STATE

STABLE RECOVERD

STATE

24

Figure 1: Time Dependent Resiliency Plot

The system functionality between 𝑡𝐸 and 𝑡𝐴 can be used as a direct measure of absorption. In

particular, the change in time with respect to functionality, i.e. the inverse of the slope, is an

intuitive measure of the system’s ability to absorb. This value can also be normalized between

zero and one, by the inverse tangent function. Equation 1 represents the system’s ability to

absorb the impact of the event. If the absorption state is 1.0, the disruption has no effect on the

system. However, a sharp, negative slope indicates poor absorption and results in value closer to

zero.

𝑅𝐴 =2

𝜋tan−1[𝜔′(𝑡)]−1 Equation 1

Similarly, the system’s ability to recover, can also be measured by the inverse of the slope within

the recovery state, 𝑡𝐷 < 𝑡 < 𝑡𝑅. Equation 2 quantifies the system’s recovery after a recovery

action has been taken.

𝑅𝑅 = 1 −[tan−1[𝜔′(𝑡)]

−1]

𝜋 Equation 2

The functionality during the disrupted state represents the system’s ability (or lack thereof) to

adapt to the adverse conditions and overcome the disruption. While system performance is no

longer decreasing, the inability to “bounce back” is measured in the disrupted state. Equation 3

provides a measure, between zero and one, for the system’s ability to quickly adapt to the new

conditions, which exist after the disruption.

𝑅𝐷 =𝑡𝐷−𝑡𝐴

𝑡𝑅−𝑡𝐸 Equation 3

Resiliency is a measure of the systems absorption (Equation 1), recovery (Equation 2), and

adaptability (Equation 3), and then a quantifiable measure of resiliency is given as Equation 4.

𝑅 = 𝑅𝐴 ∗ 𝑅𝐷 ∗ 𝑅𝑅 Equation 4

This formulation of resiliency suggests that if the system is unable to absorb or adapt or recover,

it is, effectively not resilient, 𝑅 = 0. Furthermore, this approach also allows for the

quantification of robustness, which is defined by NSF as “the loss of service that is induced by a

disturbance”1. The fractional area of system functionality between the disruption and recovery is

therefore a direct measure of the robustness of the system and provided in Equation 5.

𝜌 =∫ 𝜔(𝑡) 𝑑𝑡

𝑡𝑅𝑡𝐸

𝜔(𝑡)∗(𝑡𝑅−𝑡𝐸) Equation 5

4.2 Hybrid Multimodal Simulation Model Development

This section will focus on the description of the modeling and simulation of the six case studies

(Port of Miami, Port Everglades, Port Canaveral, Port of Jacksonville, Port of Savannah, and

Port of Charleston). In order to get a detailed view of the port operations from both the landside

and the waterside, in emergency conditions, the micro-simulation tool VISSIM was selected.

VISSIM provides the opportunity to simulate a 30-day period while being able to visually depict

25

the port operations. Along with the visual representation, this tool provides a vast array of

performance measures that can describe in detail how the ports are affected by emergency events

and their level of resilience. These models are developed from a blank slate and the applicable

available data were used to calibrate them. Furthermore, this method was used to ensure the

models represented real-life conditions. This technique will be explained further in the

methodology.

For the simulation of each port, a simulation period of thirty days was selected, in order to have a

clear overview of the whole month in which Hurricane Matthew landed. The other simulation

parameters such as the number of random seeds and runs, which are both one for every

simulation, were determined for a total of thirty simulations. The random seeds and simulation

runs have a value of one because each seed has a separate simulation run with the inputs for the

waterside arrival times. The performance measures that were used for the evaluation of the

model are the vehicle network performance and the data collection using data collection points

for each vessel type. The vehicle network performance gives an overview of both the landside

and the waterside network, while the data collection points at the entrance and exit of the port are

used in order to collect only waterside data. The aforementioned data was then extracted from

the model every hour. The results given in the report are based on an hourly evaluation of the

network.

Figure 2: Hybrid Simulation Modeling Methodology

26

4.3 Landside Model Development in VISSIM

The traffic network of each case study was modeled, and emphasis was placed on the traffic

network inside the port and the main transportation corridors that connected to its entrances and

exits. To provide more reliable and relevant results, only the roads that are directly connected to

the container and passenger terminals have been modeled. The number of lanes, port

configuration, and signal timing match that of the actual transportation network and were

incorporated into every model. These network attributes were based on Bing Maps built into

VISSIM, Google Maps, and DOT data.

Furthermore, all the models that have been developed have static routes, which were built

according to the state Departments of Transportation (DOT) traffic data and turning movement

counts. Using static vehicle routes, we assume that the turning movements and the routes will be

the same for every simulation. The car and truck volumes entering the network were extracted

from state DOT data available to the public. The inputs were calculated according to the Average

Annual Daily Traffic (AADT) for every corridor separately.

The signal control type used in all the models is the Ring Barrier Controller (RBC) emulator,

which can simulate actuated control in a VISSIM model. The features that were edited in RBC

were the signal timings, the minimum, maximum recalls, sequence, cycle length, and vehicle

detector calls. The signal timing data was gathered from similar intersections (same number of

lanes, same configuration) that were found on the City of Boca Raton website.

Additional network attributes were added in the model in order to achieve accuracy in the

simulation. These attributes include vehicle speeds, which were set on an average speed of 50

km/h for every type of vehicle, reduced speed areas on the left and right turns, and vehicle

detectors that coordinate with each signal head on every intersection.

For the waterside models an average speed of 10 knots was used, which accounts for transit

speeds as well as a speed reduction during mooring and un-mooring evolutions. In the models

pictured below, the dark grey road (links) represent the ship traffic routes in and out of port for

each of the four vessel types studied. Also pictured below are the queue of vessels waiting to

enter each port during the respective port closures. A “dummy” public transit line was created for

each ship type, 4 in total, to block the flow of ships entering the port during the entirety of the

closures

Port of Miami

For the case study of Port of Miami (Figure 4), the main roads that were modeled along with the

side roads are Biscayne Boulevard on the west side of the port, Port Boulevard connecting the

port with the mainland, MacArthur Causeway inside the port, and Port of Miami Tunnel

connecting Watson Island to Dodge Island. The entrances and exits of the port that were included

in the model are the ones from Port Boulevard and from Port of Miami Tunnel.

27

Figure 3: Port of Miami

Port Everglades

The main roads that were modeled for Port Everglades (Figure 5) are A1A on the North side of

the port, US1 on the West side of the port, Port Everglades Expressway connecting the port to

the Fort Lauderdale-Hollywood International Airport, and Eisenhower Boulevard inside the port.

Except the main roads mentioned, all the side roads that connect the main corridors with the

terminals are modeled. All the entrances and exits of the port were drawn including the ones on

the intersections of A1A and Eisenhower Boulevard, US1 and East State Road 84, and Port

Everglades Expressway and McIntosh Road.

Figure 4: Port Everglades

28

Port Canaveral

The main corridors that are included in the Port Canaveral model (Figure 6) are A1A on the west

side of the port, 401, and George King Boulevard inside the port. Additionally, more links were

created for all the side roads that connect the terminals to the main corridors. The main entrance

of the port from A1A was modeled. Vessels entering and leaving ports were modeled in VISSIM

(Figure 6a).

Figure 5: Port Canaveral

Figure 6a: Port Canaveral Closed with Vessels Queueing

Port of Jacksonville

Since the terminals of the Port of Jacksonville (Figure 7) span along the sides of St. Johns River

the terminals in this model are connected through the waterside. The main roads that were

modeled in the case of Port of Jacksonville are Heckscher Dr. 105 that connects the Blount

Island Marine Terminal with the Dames Point Marine Terminal, and E 21st St. that connects the

Talleyrand Marine Terminal to the main corridors adjacent to it. All the entrances and exits of

29

the port, including the main and side roads, were added to the model. Screen view of an example

of the simulation vessels queuing outside the port is depicted in Figure 7a.

Figure 6: Port of Jacksonville

Figure 7a: Modeled vessels queuing at closed Port of Jacksonville

Port of Savannah

West Bay St. and Main St. are the main corridors that extend across the port Savannah, along the

Savannah River, and are part of Port of Savannah model. All the side streets that connect the port

terminals with the aforementioned roads were drawn as a part of the whole network. Figure 8a

depicts detailed view of a modeled sample road junction in the road network.

30

Figure 7: Port of Savannah

Figure 8a: Port of Savannah Example Landside Intersection Model

Port of Charleston

The simulation model of the Port of Charleston (Figure 9) is mainly based on the waterside, due

to the configuration of the port. The terminals are located along the Coast of Cooper River, in

different locations. The main road that was modeled is the Mark Clark Expressway (I526) that

connects Wando Welch Terminal the rest of the terminals. For the terminals of North Charleston,

Columbus, Union Pier, and Veterans are connected through the waterside and the vessel routes.

Detailed view of a modeled sample road junction is shown in figure 9a.

31

Figure 8: Port of Charleston

Figure 9a: Port of Charleston Example Landside Intersection Model

4.4 Waterside Model Development

This section discusses the port waterside simulation model development of six case study ports:

Port of Miami, Port Everglades, Port Canaveral, Port Jacksonville, Port Savanna, and Port

Charleston. A waterside simulation was developed for each of these ports and then intergraded

with the landside model within VISSIM. Waterside operations were simulation using the Monte

Carlo approach. Fundamentally, Monte Carlo simulations work by estimating unknown values

from probability distribution functions. Effectively, vessel arrivals and departures are random

variables, however, they can be accurately estimated if enough observations are available.

Therefore, using the Monte Carlo approach to simulation modeling, it is possible to generate

random arrivals and dwell times that statistically match observed port operations. In effect this

procedure allows for the generation of a “typical” day, week, or month of vessel traffic that is

statistically similar to reality.

32

The Monte Carlo simulation approach required extensive data collection and processing to

identify the underlining statistical distributions which represent the port system. The vessel data

were analyzed to identify probability distributions of vessel arrivals and dwell time by cargo type

and time of day. From this analysis it was possible to determine, with quantifiable accuracy, the

probability of a vessel arrival, its dwell time, and its cargo type. For example, the Monte Carlo

simulation estimates the likelihood of a container vessel arriving between 7:00 and 8:00 AM and

dwelling for 25 to 26 hours before departing. From this information, the Monte Carlo simulation

generates random arrivals and dwell times that correspond to the observed probabilities. Vessel

arrivals (if and when a vessel is generated in the network) and dwell times (how long each vessel

occupied the berth), was modeled through a process called random inverse sampling. The results

of the Monte Carlo simulation were evaluated both for goodness-of-fit qualitatively, by plotting

simulated distributions alongside observed and quantitatively using regression analysis.

To develop the Monte Carlo simulation of the six ports, 12 months of AIS data from each port

was purchased from MarineTraffic.com to develop the needed arrival and departure

distributions. Vessels were categorized into four broad groups: 1) Containerized Cargo, 2) Non-

containerized Cargo, 3) Tankers, and 4) Passenger Ships. These categories were chosen for their

pervasiveness at each port and the unique loading and unloading characteristics of the cargo they

carry. The data contained 71,795 records of vessel arrivals, departures, and dwell time, starting

January 1st, 2016 and ending December 31th, 2016. Three additional months of data was also

purchased for each port for model validation purposes. This data encompassed January 1st, 2017

through March 31st, 2017.

Table 1 summarizes the number arrivals, by category within the study ports. From the table it can

be seen that most of the vessels carried containerized cargo. These four vessel categories make

majority of cargo vessel observations within the study ports. For this reason, the simulation

models were focused on only these four vessel categories.

Table 1: Cargo Type by Study Port

NUMBER OF VESSELS THAT ARRIVED BETWEEN 1/1/2016 & 12/31/2016

MIAMI EVERGLADES CANAVERAL CHARLESTON JACKSONVILLE PALM BEACH SAVANNAH

CONTAINER 980 1908 134 3103 2228 819 3943

NON-CONT. 613 474 231 417 557 156 149

TANKER - 338 205 270 327 - 336

PASSENGER 1055 778 2646 277 205 215 1572

With the data stratified by cargo type, frequency distribution functions were developed for vessel

arrivals and vessel dwell time. Vessel arrivals were categorized into one-hour bins, i.e., if a

vessel arrived at 4:25 AM, then it was counted in the 4:00 AM – 5:00 AM bin. Vessel dwell

times were also categorized into one-hour bins, i.e. if a vessel dwelled for two hours and forty-

five minutes, then it was counted in the two three-hour bin. This approach was taken for all cargo

types and all hours of the day. The next step was to generate the probability distribution function

(PDF) by dividing the frequency distribution by the total number of observed vessels. The

cumulative distribution function (CDF) is the integral of the PDF and was generated for vessel

33

arrival time and dwell time. Figure 9 and

Figure 10 show the arrival and dwell time PDF and CDF for container vessels at Port

Everglades, respectively. The figure shows the arrivals to be more or less uniformly distributed.

Dwell times for container vessels at the Port Everglades appear to be a right skewed, normal

distribution. The arrival CDF appears to be linear, again suggesting a uniform distribution and

the dwell time CDF displayed a distinctive “S-curve”, indicative of a normal distribution.

The arrival and dwell time PDFs are located in APPENDIX A and are displayed alongside the

simulated values and values obtained from the validation dataset. In general, vessel arrivals were

uniformly distributed throughout the day. However, passenger arrivals were normally distributed

within a morning peak period. The distribution of vessel dwell times tended to be unique within

each cargo type. Tanker and non-containerized cargo vessels appeared to have a uniformly

distributed dwell time. Container vessels tended to be right-tailed normal distributions and

passenger vessels appear normally distributed with means in the range of nine to 13 hours.

0

0.2

0.4

0.6

0.8

1

0%

1%

2%

3%

4%

5%

6%

7%

8%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME (HOURS)

PORT EVERGLADES CONTAINERIZED CARGO VESSEL DWELL TIME (PDF & CDF)

PROBABILITY DISTRIBUTION FUNCTION CUMULATIVE DISTRIBUTION FUNCTION

34

Figure 9: Port Everglades Container Vessel Dwell Time PDF

Figure 10: Port Everglades Container Vessel Arrival PDF

Vessel arrivals and dwell times were then simulated using the Inverse Transform Sampling

method. Fundamentally, this approach uses the inverse of the CDF to transform a uniformly

distributed random value into an arrival time or dwell time that matches the historic distribution.

Using the cumulative distribution function from containerized vessels at the Port of Jacksonville

as an example in Figure 11, one-hour vessel arrival bins are shown on the x-axis and the

cumulative probability (between zero and one) is shown on the y-axis. Taking the inverse of this

function transforms a uniformly distributed random number, through the line shown in the figure

and matches it up to a one-hour arrival bin. Effectively, this would be like pointing a finger at a

0

0.2

0.4

0.6

0.8

1

0%

1%

2%

3%

4%

5%

6%

7%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

TIME OF DAY (HH:MM)

PORT EVERGLADES CONTAINERIZED CARGO VESSEL ARRIVAL BY TIME OF DAY (PDF & CDF)

PROBABILITY DISTRIBUTION FUNCTION CUMULATIVE DISTRIBUTION FUNCTION

0

0.2

0.4

0.6

0.8

1

0%

1%

2%

3%

4%

5%

6%

7%

8%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME (HOURS)

PORT EVERGLADES CONTAINERIZED CARGO VESSEL DWELL TIME (PDF & CDF)

PROBABILITY DISTRIBUTION FUNCTION CUMULATIVE DISTRIBUTION FUNCTION

35

random value on the y-axis and dragging it toward the CDF line to find the corresponding arrival

bin on the x-axis. Figure 12 shows an example of a randomly selected value of 0.7 from a

uniform distribution being transformed to a randomly assigned arrival time of approximately

6:00 PM. In this fashion, it was possible to generate random arrivals times that match the

observed distributions. Applying the Inverse Transform Sampling Method, vessels arrivals were

simulated for 10, one-year periods. This was done to verify or calibrate the vessel arrival model.

Once confirmed, a 30-day period was also generated and integrated into the VISSIM simulation

of landside operations.

Figure 11: Inverse Transform Sampling Example: Port of Jacksonville Containerized Vessel CDF

Vessel arrivals were considered a discrete variable falling into any one of 24 time bins. Once a

vessel was assigned an hour to arrive, the vessel was projected to enter the port either at the hour,

or 15-minutes, 30-minutes, or 45-minutes past the hour. This was assigned on a random basis

from a uniform distribution. Vessel dwell times on the other hand are a continues variable and

unlike arrivals which only have 24 possible bins, the number of one-hour dwell time bins

possible nearly is infinite. Therefore, when developing the dwell time distributions, the number

of bins was limited to include most, but not all observed dwell times. Furthermore, because it

was desirable for the model to produce a continuous variable for the dwell time, the modeling

approach for dwell times was slightly modified. The CDF of the dwell time was approximated to

be a linear function. The inverse of this linear approximation was sampled from to generate the

dwell times in a similar fashion as the arrival times. This approach to modeling dwell times lead

to in less accurate results than other possible methods because of the linearity assumption but

was computation efficient. Table 2 shows the linearized inverse sampling questions for vessel

dwell times by port and cargo type.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0:00 4:48 9:36 14:24 19:12 0:00 4:48

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF JACKSONVILLE ARRIVAL CUMLATIVE DISTRIBUTION FUNCTION: CONTAINERIZED CARGO VESSELS

36

Table 2: Dwell Time Inverse Random Sampling Equations

4.6 Integrating the Water and Landside Simulation Models

The next step in the development of the simulation models was to incorporate the vessel arrivals

which are able to accurately depict the port operations during a major weather event. For this

process, a simulated 30-day schedule of port operations was developed for five random seeds or

trials for each port. Within VISSIM, vessels were modeled as public transport vehicles to

represent sea traffic. Public transport in VISSIM provides attributes that were beneficial in

creating the waterside network of the ports. These attributes include departure times, dwell times,

routes and stops (berths) for each public transport line.

The procedure that was followed in VISSIM for the waterside model development started by

creating the links (paths) for the vessels to travel to and from the port. In order to get more

accurate results for each port, only the paths that connect the berths with the entrance/exit of the

port were drawn. In each model, the paths were designed according to the geometry of the port

and the information of the existing vessel traffic patterns. The link modeling was followed by the

establishment of the public transport stops, which represent the port berths. Each public transport

stop has the length and the position of the corresponding berth. After the public transport stops,

the public transport lines were created. Each public transport line connects to one public

transport stop (berth), creating a specific route for every vessel that is destined to that berth.

Additionally, each public transport line berth has a specific vehicle type (vessel type). Also, the

speeds of all the vessels were varied depending on the traffic patterns and speed restrictions of

each specific port.

The simulated vessel arrivals were modeled in VISSIM as public transport departure times.

Vessel dwell times were modeled as occupants. The time required for each occupant to exit the

public transit vehicle was deterministically set to 999 seconds. The simulate dwell times were

37

converted to 999 second units and programed as occupants. When the vessel arrives at the berth,

it dwells while the “occupants” exits the vessel. Hence, by using public transport lines VISSIM

was able to model any type of vessel (container, non-container, tanker, and passenger) to the

correct berth at the correct arrival and dwell time.

4.7 Scenario Development

Three simulated scenarios were developed as part of this research. The first scenario represents a

typical 30-day period of port operations beginning midnight of September 29th, 2016 and ending

at 11:59:59 on October 28th, 2016. This scenario represented the baseline case and was titled as

such; Baseline Case. The baseline case scenario established a bench mark for port operations

with which to compare subsequent results with. The baseline case scenario results, like all of the

modeled scenarios, constitutes five simulated trials or “runs”. When conducting stochastic

simulations, runs are needed to rule out the effects of randomness in the model which may

impact the research findings. For example, if a fair dice is rolled twice and the result is 2 each

time, someone could come to the false conclusion the dice is weighted. However, as the number

of runs increase, the likelihood of coming to a false conclusion by chance alone, is reduced.

The second scenario represents the impact of Hurricane Matthew and is titled Hurricane

Matthew. Similar to the baseline case scenario, the simulation begins on September 29th, 2016

and has the same five simulation runs as the baseline case (same arrival and dwell pattern). In

this scenario, all data for the closure and re-opening of the ports was obtained from the Marine

Safety Information Broadcast messages published by the Coast Guard and that data was modeled

into each port. Table 3 shows the port closures and reopening times programed into the

simulation. During the closure, access to the port was prohibited. This resulted in a queue of

vessel waiting to enter the port. While no such queue exited during Hurricane Matthew (vessel

did not wait just outside the port during the storm), it represents a backlog of vessels which

would have entered the port, if not for the storm and closure. This scenario also assumed that all

vessel in a typical month would call their respective ports and not re-route to do the hurricane.

Once each port was reopened in the simulation, the vessel backlog was serviced, resulting in

increased demand placed on port infrastructure. Additionally, each vessel was required to wait

until berthing space became available. From a modeling perspective, the port was closed by

creating a “dummy vessel”. These dummy vessels were set to arrival and occupy the entrance of

the port for the duration of the closure. Once the port was reopened, the vessel exited, and port

operations resumed. It is noted that in the cases considered here, each port is assumed to be

operating at 100% efficiency to service the queue of ships once it is opened. Other situations can

be considered as required.

A third scenario was developed to model each port with additional berths for each vessel type

after the port opened. The amount of additional available berth space varied between ports and

between vessel types. On average additional 2-3 vessels were allowed to be serviced at the same

time at a single berth. Allowing additional “hypothetical” berthing space can represent a

multitude of real-life scenarios that can greatly help port partners improve their resiliency both

on local and regional scales. The most obvious conclusion that can be drawn from this scenario

is what infrastructure upgrades can impact the resiliency of a port. Whether it is additional

berthing space or, where no more land exists to develop, additional equipment to expedite the

onboarding and offboarding of cargo. This scenario can also represent the impact of increasing

38

the manpower to improve the efficiency of cargo operations. Finally, this scenario involving

extended berthing space can also represent ship-to-ship transfer of cargo in order to reduce the

queue. For example, it can be used to assess the impact of tanker vessels conducting lightering

operations at a tanker barge. These results can assist port partners with pre-storm planning, more

specifically potentially pre-staging tanker barges to expedite the clearing of the tanker vessel

queue. Multiple results from a scenario of this type can be considered, in support of increasing

the resiliency of the ports considered. This scenario is titled Capacity Enhanced in the results in

section 5.

Table 3: Port Closures Resulting from Hurricane Matthew

PORT CLOSURE

TIME 2016

REOPEN

TIME 2016

HOURS CLOSED

ESTIMATED ECONOMIC

LOSS

(Millions of $)

MIAMI 10/5 22:00 10/7 10:00 36 $169 M

EVERGLADES 10/5 22:00 10/7 12:00 38 $129.2 M

CANAVERAL 10/5 22:00 10/9 8:00 82 $20.5 M

JACKSONVILLE

PASSENGER BERTHS

10/6 8:00 10/8 23:00 63

JACKSONVILLE ALL

BERTHS

10/6 8:00 10/9 9:00 73 $224 M

SAVANNAH 10/7 8:00 10/12 7:00 120 $600 M

CHARLESTON 10/7 16:00 10/10 5:00 61 $372.1 M

TOTAL 410 $1.51 Billion

The last column in Table 3 shows estimated regional economic loss due to the port

closures during Hurricane Matthew, based on published data on the contribution each port makes

to the local economy and the period of port closure. The estimates do not account for loss

revenues for the ships waiting to offload cargo or the cruise ships that had to cancel or change

routes. The port closure data was obtained through the Maritime Safety Information Broadcasts

published by the United States Coast Guard.

(Sources:https://www.miamidade.gov/portmiami/,http://www.porteverglades.net/,

https://www.portcanaveral.com/, https://www.jaxport.com/, http://gaports.com/port-of-savannah,

http://www.scspa.com/)

39

5.0 RESULTS

In this section, the model calibration and validation results are discussed and the results of the

vessel activity for the baseline cases and for the studied scenarios are analyzed and compared.

The calibration and validation provide a measure of how well the model preformed when

compared to expectations.

5.1 Model Calibration and Validation

Four separate datasets were used in the calibration and validation of the simulation model. The

first was a 12-month dataset acquired to generate the simulation models. This dataset was

referred to as 12-Month Cal (calibration) dataset. Next, the simulation model was used to

generate ten, 1-year periods. This allowed for the model to generation enough observations to

verify the accuracy of simulation predictions. This dataset was titled 10 Year Sim (simulation).

Next, five 30-day simulation runs were developed and evaluated for their accuracy. This dataset

was titled 30 Day Sim (simulation). Finally, the validation dataset purchased from

MarineTraffic.com was also shown to better validate the model findings. The validation dataset

was titled 3 Month Val (validation).

The model calibration and validation were conducted in two stages. First, a qualitative

assessment was investigated to provide a consensus of how the models preformed. This was

accomplished by plotting the historical probability distribution functions (12 Month Cal. and 3

Month Val.) alongside the distribution functions generated by the simulation model (10-Year

simulation and 30-Day simulation). Figure shows the arrival distribution functions (historic and

simulated) for container vessels at the Port of Miami. The figure suggests a general consistency

between the four data sources. Appendix A: Arrival and Dwell Time Distribution Functions,

contains similar figures for each of the four vessel types at all six case study ports. In general,

distribution plots with a higher number of observations and more consistent 12-month historical

distributions, showed more consistent model and validation results. Likewise, distribution built

upon fewer observations or ones with inconsistent observations made during the 12 historical

dataset, produced less stable models results. Overall, the plotted distributions represented the

essence of the operations at each port. While dwell times tended to under preform, this was

expected because of the linearity assumption made in the modeling process. Furthermore,

predicting which hour a vessel would arrive is fundamentally easier than predicting how many

hours it is likely to remain within the port.

40

Figure 13: Port of Miami Containerized Cargo Arrival Time Distributions

The second calibration and validation approach used the coefficient of determination, also

known as R-Squared value, to quantitatively evaluate the Goodness-of-fit. Goodness-of-fit refers

to how well a simulation model matches the observed values. The coefficient of determination is

a measure of how well the model’s regression line estimates the real data points. A coefficient

value of one indicates a perfect fit. The coefficient of determination is a strict measure of model

performance. Vessels arriving later than expected not only impact the current time period but the

following time period as well and necessarily result in an R-Squared value less than 1.0.

For the purpose of this study, the calibration results were estimated by the coefficient of

determination when comparing the 12 Month Cal dataset to both the 10 Year simulation and the

30-Day simulation. This was done for vessel arrival time in Table 4 and vessel dwell time in

Table 5. The results shown in Table 4 suggest accurate results with regard to when vessel entered

the port. R-Squared values ranged between a low of 0.302 and a high of 0.999, with the majority

of the results above 0.90. Table 5 shows a range of dwell time R-Squared values between 0.001

and 0.516. This suggest the model struggled to reproduce the exact dwell times seen in the 12

Month Cal dataset. This was expected due to simplified modeling approach taken for dwell times

and the assumption of a linear CPF. However, when looking in context of the distribution plots

and the strict criteria of the R-Squared measure, the model represented port operations

reasonably well for the purpose of the scenario comparisons.

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF MIAMI ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

41

Table 4: Arrival Time Calibration Results

CONTAINER NON-CONTAIN. TANKER PASSENGER

PORT 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY

CHARLESTON 0.998 0.916 0.960 0.358 0.857 0.176 0.995 0.828

SAVANNAH 0.997 0.824 0.985 0.754 0.942 0.478 1.000 0.981

JACKSONVILLE 0.995 0.821 0.941 0.581 0.951 0.302 0.999 0.936

CANAVERAL 0.940 0.376 0.955 0.540 0.929 0.381 1.000 0.988

EVERGLADES 0.986 0.485 0.970 0.476 0.974 0.422 0.997 0.947

MIAMI 0.998 0.704 0.994 0.706 N/A N/A 1.000 0.078

Table 5: Dwell Time Calibration Results

CONTAINER NON-CONTAIN. TANKER PASS

PORT 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY

CHARLESTON 0.370 0.473 0.012 0.063 0.150 0.033 0.287 0.249

SAVANNAH 0.314 0.350 0.025 0.013 0.003 0.008 0.099 0.104

JACKSONVILLE 0.441 0.341 0.308 0.069 0.404 0.132 0.181 0.595

CANAVERAL 0.106 0.018 0.017 0.002 0.158 0.154 0.469 0.516

EVERGLADES 0.271 0.114 0.170 0.221 0.499 0.310 0.248 0.184

MIAMI 0.002 0.001 0.042 0.007 N/A N/A 0.370 0.212

For the purpose this research, validation was measured by the R-Squared value generated by the

comparison of the simulation datasets and the 3 Month Val dataset. In general, the R-Squared

values found during validation are lower when compared to calibration. Table 6 shows the

validation results for the arrival of vessels. The R-Squared value ranged between 0.001 and

0.729. One reason for the lower R-Squared values could be that the validation dataset was taken

over a consecutive three-month period. The simulation models developed did not account for

seasonable variability and therefore if any of the ports experience seasonal changes in cargo

shipments, this would not be reflected in the model. Table 7 7 shows the dwell time validation

results. The R-Squared value ranged between 0.001 and 0.472. Again, these R-Squared values

were reflective of the modeling approach, assumptions, and likely the seasonal impact of port

operations. And while these values tend to be lower, the observed distributions, provided in

Appendix A, show a general agreement between the model results and the historical datasets.

42

Table 6: Arrival Time Validation Results

CONTAINER NON-CONTAIN. TANKER PASSENGER

PORT 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY

CHARLESTON 0.662 0.630 0.028 0.168 0.007 0.148 0.510 0.394

SAVANNAH 0.729 0.624 0.041 0.050 0.005 0.004 0.426 0.451

JACKSONVILLE 0.584 0.413 0.024 0.105 0.095 0.113 0.317 0.405

CANAVERAL 0.001 0.005 0.072 0.007 0.044 0.037 0.556 0.571

EVERGLADES 0.355 0.045 0.036 0.021 0.331 0.142 0.639 0.593

MIAMI 0.021 0.001 0.354 0.372 N/A N/A 0.421 0.007

Table 7: Dwell Time Validation Results

CONTAINER NON-CONTAIN. TANKER PASSENGER

PORT 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY 10 YR 30 DAY

CHARLESTON 0.321 0.387 0.003 0.000 0.028 0.004 0.215 0.184

SAVANNAH 0.300 0.255 0.001 0.001 0.000 0.001 0.086 0.066

JACKSONVILLE 0.383 0.325 0.018 0.000 0.073 0.082 0.003 0.001

CANAVERAL 0.054 0.009 0.001 0.001 0.033 0.010 0.420 0.472

EVERGLADES 0.173 0.027 0.093 0.041 0.392 0.259 0.240 0.153

MIAMI NULL NULL 0.014 0.003 N/A N/A 0.034 0.174

5.2 Scenario Results

The simulation results from VISSIM provide a wide range of output parameters. The key

measure of effectiveness (MOE) selected to present here was port occupancy. Port occupancy

was defined for this research as the number of vessels of a particular cargo type occupying the

port at given moment in time. In this sense, occupancy provides an indication of port capacity.

The integration of the waterside and landside models within VISSIM allowed for vessels to

interact in both time and space. Therefore, as port becomes congested after the closure, capacity

and by extension occupancy become a significant concern for port managers. The port

occupancy was determined by analyzing vessel entries and exits.

In general, the most frequent cargo type observed at the case study ports was containerized

cargo. This was true for all, but one, port: Port Canaveral. The latter experiences its high traffic

from passenger vessels. Because the analysis of port capacity was the primary consideration only

the cargo types with the highest demand at each port are described here. A preliminary analysis

of the lower demand cargo types found little impact to port operations caused by the storm

closures. This was because, in general, these vessels were easily serviced following the port

reopening.

Figure shows the containerized cargo vessel occupancy for the Port of Miami. The orange line

represents the baseline case scenario, in the absence of the disruption. The blue line shows the

modeled port operations during Hurricane Matthew and the black line represents the scenario

43

involving enhanced port capacity. In general, the impact of the closure on containerized cargo

vessels at the Port of Miami was minor. This is to be expected because Miami was located

furthest from storm’s path and the port experienced as shorter closure because of this. By

integrating the difference between the Hurricane Matthew scenario and the baseline case, an

estimate of the total number of lost vessel-hours at Port of Miami was obtained. On average, the

simulation model estimated a total of 48 vessel-hours of lost time due to the disruption.

Furthermore, it does not appear that capacity enhancements had any meaningful impact on port

operations. Again, this was likely because the port was out of the storms direct path and only

experienced a minimal disruption.

Figure shows containerized vessel occupancy for Port Everglades. This figure follows the same

convention as the prior one. The figure shows a more significant impact on Port Everglades. This

was likely because, in general, Port Everglades processes more containerized cargo when

compared to the Port of Miami. Port Everglades experienced an estimated 134 vessel-hours of

lost time, on average as a result of the closure. However, this was relatively minor as Port

Everglades was not directly impacted by the storm and the resulting closure was limited.

Figure shows that the passenger vessel occupancy results for Port Canaveral. It follows the same

pattern as the prior two figures. The figure suggests a significantly larger backlog of vessels

when the port was reopened. While it is unlikely that passenger vessels would queue in such a

way, the analysis does provide for a theoretical examination of the passenger vessel capacity at

the port. The total number of vessel-hours lost due to the closure was estimated at around 586.

This is nearly eight times the impact felt at Port Everglades. The capacity enhancements

provided in scenario three were estimated to reduce the loss by approximately 245 vessel-hours

through servicing the queue of vessels more efficiently after the reopening.

Figure shows the containerized cargo vessel occupancy at the Port of Jacksonville. The figure

clearly shows a significant impact of the closure on the port as well as extended duration of time

needed to address the backlog of cargo vessels. The impact of the closure was felt at this port

beginning October 6, 2016 at 1:00 PM and the port did not return to a normal state until October

21, 2016 at 9:00 AM, over two weeks later. The total lost vessel time was estimated at

approximately 3,200 vessel-hours. The modeled capacity enhancement case suggested a possible

saving of approximately 2,500 vessel-hours as a result of the enhancement.

Figure 18 shows the containerized cargo vessel occupancy at the Port of Savannah. The figure

again shows a significant impact of the closure on the port as well as extended duration of time

needed to address the backlog of cargo vessels. The impact of the closure was felt at this port

beginning October 7, 2016 at 6:00 PM and the did not return to a normal state until the end of the

simulation run on October 28, 2016. The total lost vessel time was estimated at over 5,000

vessel-hours. The capacity enhancements showed insignificant change.

Figure 19 shows the containerized cargo vessel occupancy at the Port of Charleston. The figure

again shows a relatively significant impact of the closure on the port as well as extended duration

of time needed to address the backlog of cargo vessels. The impact of the closure was felt at this

port beginning October 7, 2016 at 7:00 PM and did not return to a normal state until the end of

44

the simulation on October 28, 2019. The total lost vessel time was estimated at over 900 vessel-

hours. The capacity enhancements showed insignificant change.

Figure 14: Port of Miami Containerized Cargo Vessel Occupancy Results

Figure 15: Port Everglades Containerized Cargo Vessel Occupancy Results

Figure 16: Port Canaveral Passenger Cargo Vessel Occupancy Results

45

Figure 17: Port of Jacksonville Containerized Cargo Vessel Occupancy Results

Figure 18: Port of Savannah Containerized Cargo Vessel Occupancy Results

Figure 19: Port of Charleston Containerized Cargo Vessel Occupancy Results

46

Figure 20 shows the passenger vessel occupancy results for the Port of Savannah. The figure

shows the significant impact caused by the closure and ensuing queue of vessel waiting to gain

entry into the port. The simulation results suggest the backlog of passenger vessel would likely

extend beyond the 30-day simulation period, if actions by port managers were not taken. In

reality, many of the passenger vessel trips were canceled or rerouted to other ports. Nevertheless,

the simulation provides an excellent example to quantify the benefits gained from the planning

and recovery actions taken during Hurricane Matthew. While the Hurricane Matthew scenario

did not converge within the simulation time, the total vessel-hours lost at the end of the

simulation was approximately 5,000. However, the capacity-enhanced scenario resulted in a total

lost time of 1,500 vessel hours, a total saving of 3,500 vessel hours.

Figure 20: Port of Savannah Passenger Vessel Occupancy Results

Figure 21 shows the non-containerized cargo vessel occupancy results from Port Charleston. In

general, the port sees only minor non-containerized vessel cargo during the 30-day simulation

period. However, the impact of the Hurricane Matthew closure was still felt. The simulation

resulted in estimated loss of 240 vessel-hours during the closure and recovery period. The results

also suggest that currently capacity levels within the port were sufficient to service the backlog

of vessels upon the reopening of the port.

The lost vessel hours at the ports are shown in Table 8 and provide a comparison of the impact of

the storm between the ports.

0

5

10

15

20

25

9/2

9/2

016

0:00

9/2

9/2

016

15:0

09/

30

/201

6 7:

009/

30

/201

6 23

:00

10/1

/201

6 15

:00

10/2

/201

6 7:

0010

/2/2

016

23:0

010

/3/2

016

15:0

010

/4/2

016

7:00

10/4

/201

6 23

:00

10/5

/201

6 15

:00

10/6

/201

6 7:

0010

/6/2

016

23:0

010

/7/2

016

15:0

010

/8/2

016

7:00

10/8

/201

6 23

:00

10/9

/201

6 15

:00

10/1

0/2

016

6:0

010

/10/

201

6 2

2:00

10/1

1/2

016

14:

0010

/12/

201

6 6

:00

10/1

2/20

16 2

2:00

10/1

3/2

016

14:

0010

/14/

201

6 6

:00

10/1

4/2

016

22:

0010

/15/

201

6 1

4:00

10/1

6/20

16 6

:00

10/1

6/2

016

22:

0010

/17/

201

6 1

4:00

10/1

8/2

016

6:0

010

/18/

201

6 2

2:00

10/1

9/2

016

14:

0010

/20/

201

6 6

:00

10/2

0/2

016

21:

0010

/21/

201

6 1

3:00

10/2

2/20

16 5

:00

10/2

2/2

016

21:

0010

/23/

201

6 1

3:00

10/2

4/2

016

5:0

010

/24/

201

6 2

1:00

10/2

5/20

16 1

3:00

10/2

6/2

016

5:0

010

/26/

201

6 2

1:00

10/2

7/2

016

13:

0010

/28/

201

6 5

:00

10/2

8/2

016

21:

00

VES

SELS

PORT OF SAVANNAH PASSENGER VESSELS OCCUPENCY

BASE CASE HURRICANE MATHEW CAPACITY ENHANCED

47

Figure 21: Port of Charleston Non-Containerized Cargo Vessel Occupancy Results

Table 8: Comparative Impact in terms of lost containerized vessel hours

Port Time

Closed

Time

Open

Hours

Closed

Container Vessel

Hours Lost

Hurricane

Matthew

Container Vessel

Hours Lost

Enhanced Capacity

MIAMI 10/5

22:00

10/7

10:00

36 48 48

EVERGLADES 10/5

22:00

10/7

12:00

38 134 134

CANAVERAL 10/5

22:00

10/9 8:00 82 586(Cruise Ship

hours)

341(Cruise Ship

hours)

JACKSONVILLE 10/6 8:00 10/9 9:00 73 3,200 700

SAVANNAH 10/7 8:00 10/12

7:00

120 5,000 5,000

CHARLESTON 10/7

16:00

10/10

5:00

61 900 900

Total 410 9,868 7,123

5.3 Regional Impact

The impact of Hurricane Matthew to the region was significant. While most of the effects were

concentrated in the northern ports, measurable losses were observed at the Ports of Miami and

Everglades. The simulated results suggest that in total it is likely that Hurricane Matthew may

have caused over 30,000 lost vessel-hours to the region, including a total of over 9,200

containerized cargo- vessel hours between the six ports considered. The analysis suggests that

without the action taken by the U.S. Coast Guard, U.S. Army Corps of Engineers, and port

managers, the adverse effects of this major disruptive event would have been significantly more

disastrous. However, the planning and recovery actions pursued in reality mitigated these

impacts. The analysis and results suggest an additional opportunity for enhanced regional

resiliency. While many of the northern ports experienced significant queues and delays, the

southern ports were operating at normal capacity within a few days of reopening. Therefore, for

cargo that can be transported over the road network it makes sense to reroute these vessels to

0

1

2

3

4

5

6

7

8

9/2

9/2

016

0:00

9/2

9/2

016

15:0

09/

30

/201

6 7:

009/

30

/201

6 23

:00

10/1

/201

6 15

:00

10/2

/201

6 7:

0010

/2/2

016

23:0

010

/3/2

016

15:0

010

/4/2

016

7:00

10/4

/201

6 23

:00

10/5

/201

6 15

:00

10/6

/201

6 7:

0010

/6/2

016

23:0

010

/7/2

016

15:0

010

/8/2

016

7:00

10/8

/201

6 23

:00

10/9

/201

6 15

:00

10/1

0/2

016

6:0

010

/10/

201

6 2

2:00

10/1

1/2

016

14:

0010

/12/

201

6 6

:00

10/1

2/2

016

22:

0010

/13/

201

6 1

4:00

10/1

4/2

016

6:0

010

/14/

201

6 2

2:00

10/1

5/2

016

14:

0010

/16/

201

6 6

:00

10/1

6/2

016

22:

0010

/17/

201

6 1

4:00

10/1

8/2

016

6:0

010

/18/

201

6 2

2:00

10/1

9/2

016

14:

0010

/20/

201

6 6

:00

10/2

0/2

016

21:

0010

/21/

201

6 1

3:00

10/2

2/2

016

5:0

010

/22/

201

6 2

1:00

10/2

3/2

016

13:

0010

/24/

201

6 5

:00

10/2

4/2

016

21:

0010

/25/

201

6 1

3:00

10/2

6/2

016

5:0

010

/26/

201

6 2

1:00

10/2

7/2

016

13:

0010

/28/

201

6 5

:00

10/2

8/2

016

21:

00

VES

SELS

BASE CASE HURRICANE MATHEW CAPACITY ENHANCED

48

neighboring ports. This would decrease the initial queue or backlog of vessels needing to be

serviced by the most significantly impacted ports. Decreasing the initial queue would likely have

an exponential impact on vessel-hours lost because the relationship between initial queue and

recover time was observed to be non-linear. This was also seen for the modeling of the enhanced

capacity scenarios. Because at Jacksonville, the port was better able to accommodate the sudden

surge in vessel traffic, port operations returned to normal sooner and therefore the cascading

impact of new vessel arrivals did not manifest.

49

6.0 DISCUSSION AND CONCLUSIONS

A modeling and simulation-based framework has been developed for determining consequences

of disruption at ports in a region due to the passage of a hurricane event. In particular,

disruptions at a group of case-study ports due to Hurricane Matthew, a category 5 Atlantic

hurricane that skirted the southeast US coast in October 2016, are considered. Significant

stakeholder engagement through a well-attended workshop, visits to ports and meetings with

USCG personnel, and other forums resulted in useful input to the development of the project.

The methodology used in developing the framework involves establishment of detailed models

of the selected regional ports and their transportation networks on the VISSIM software

platform, which together with custom Monte Carlo-based hybrid optimization algorithms

integrated into the platforms, as well as libraries of algorithms already built into the platforms,

serve as a tool for assessing and planning for the resiliency of regional ports to a hurricane event.

The tool analyzes available data on the ports and their normal baseline patterns of operation,

including arrivals and departures of intermodal transportation, to predict consequences of a

disruptive event. It provides opportunities for exploring the consequences of alternative decisions

in responding to a hurricane event. Since each port is unique, the tool is customized with specific

models to serve the given port. The general framework for the tool has been established, and a

customized tool has been developed for each of the six ports considered: Port of Miami, Port

Everglades, Port Canaveral, Port of Jacksonville, Port of Savannah, and Port of Charleston.

Baseline operations at the six case-study ports over a 30-day window including the Hurricane-

Matthew event were modeled and the impact on normal level of service at these ports due to the

disruption caused by the hurricane were evaluated. The results are used to quantify the impact on

the ports due to the event and make comparison with available empirical data for the event. It is

found that the agreement between the predicted consequences and empirical observations over

the event is good where significant input data are available at the individual ports and not so

good where limited data are available. The framework can be similarly used to develop a custom

tool for any regional port, and similar case studies can be undertaken to answer stakeholder

questions relating to the waterside and landside resiliency of a given port.

In general, the results of the research show the benefits of quantifying the impact of an event and

how the information gained from such analysis can be beneficial when evaluating alternatives. In

the case studies considered, the impact of enhanced service capacities at ports in clearing

backlogs was quantified. Such quantitative assessments provide meaning, context, and relevance

to port stakeholders that may not be readily apparent at face value. This research also shows that

Automatic Identification System (AIS) data could be utilized to create new methods and metrics

for the assessment of resiliency in maritime systems. This methodology advances the field of

disaster science by expanding on the concepts first proposed by Henry and Ramirez-Marquez

(2012) and Baroud et al. (2014) and applying these methods to empirical observations. AIS is an

excellent source for quantitative data when seeking post-disaster measures of resiliency. The

time dependent performance models developed from this data show the cascading effects of

disruptions and quantify the benefits gained by recovery efforts in a time-progressive series. The

data show, in quantifiable terms, reductions in performance resulting from the simulated

disruption. On a broad level, these findings also represent some of the first steps toward the

development of standardized metrics for quantifying MTS operational resiliency. The use of AIS

data, which collects information from commercial vessels on a semi-continuous basis, is a rich

data source with many applications in disaster science. The methods developed and applied here

50

quantifies resiliency of regional ports to a hurricane event and can be applied across a range of

temporal and spatial scales.

The case studies conducted for the six ports for the disruptive Hurricane-Matthew scenario

suggest that the modeling and simulation-based approach considered here is good for predicting

the consequences of the related disruptions at the ports and the region, provided sufficient input

data about the normal port operations are available. The developed tools allow prediction of

possible consequences of making a particular decision from a set of alternative decisions in

responding to a disruption and developing an optimal response.

6.1 Plan for Transition

The plan for transition of the developed framework and tools have been discussed with USCG

personnel. As described above, the framework for tools specific to a given regional port for

determining the consequences of a disruption to the port’s activities due to a hurricane event has

been developed. It is based on the establishment of a detailed model of the port and its

transportation network on the VISSIM software platform that includes built-in libraries of

algorithms for simulating intermodal transportation, and incorporation of custom Monte Carlo-

based optimization algorithms onto the platform. Specific tools have been developed for the six

selected ports for the case study, in support of assessing and planning for resiliency at these

ports. The tools developed can be used to play out various scenarios of interest to USCG and

other stakeholders. It has been agreed to house the framework and developed tools within the

newly established USDOT supported Freight Mobility Research Institute at FAU that can:

– Provide access and service to stakeholders for predicting consequences of a

disruption at any given Port and planning for mitigation of impact of a disruption

– Nurture further development of the framework and associated tools.

– Support other complementary DHS-funded efforts in Port resiliency

The framework and the tools can be improved and further developed, including:

– Refinements to event modeling and simulation to enable consideration of different

levels of threats

– Collection of pertinent port disruption data related to recent hurricanes (Matthew,

Irma) to support further validation of the developed framework

– Collection of additional port operations data to support improvements to micro-

level modeling of intermodal transportation within ports

A manual for running the model is being put together.

51

7.0 ACKNOWLEDGEMENTS

We are grateful to USCG LCDR Stryker (USCG POC and Project Champion) for providing

valuable feedback during periodic conference call meetings and for serving on a panel at the

stakeholder workshop in March 2019. The project has benefitted significantly from participation

by the following students:

• At FAU: Panagiota Goulianou, LT Maxwell Walker, USCG and LT Joshua Villafane,

USCG

• At ERAU: Hannah Thomas, Emily Jannace, and Fanny Kristiansson

The students assisted with running the simulations and with engaging the ports.

We are also grateful to Dr. Hady Salloum and Beth DeFares for supporting the project and for

connecting us with project stakeholders.

52

8.0 REFERENCES

1. Rose, A., & Liao, S. Y. (2005). Modeling regional economic resilience to disasters: A

computable general equilibrium analysis of water service disruptions. Journal of Regional

Science, 45(1), 75-112.

2. Mansouri, M., Nilchiani, R., & Mostashari, A. (2010). A policy making framework for

resilient port infrastructure systems. Marine Policy, 34(6), 1125-1134.

3. Nair, R., Avetisyan, H., & Miller-Hooks, E. (2010). Resilience framework for ports and

other intermodal components. Transportation Research Record: Journal of the

Transportation Research Board, (2166), 54-65.

4. Miller-Hooks, E., Zhang, X., & Faturechi, R. (2012). Measuring and maximizing

resilience of freight transportation networks. Computers & Operations Research, 39(7),

1633-1643.

5. Orosz, Michael D., "PortSec: Port Security Risk and Resource Management System"

(2011). Research Project Summaries. Paper 87.

6. White Paper - "The MIT CTL Port Resilience Survey Report" by James B. Rice, Jr. and

Kai Trepte, December 2012

7. "Port Mapper: Preparing for the Future" by James B. Rice, Jr., Kai Trepte and Ken

Cottrill, The Eno Brief, June 2013

8. Babun, T., & Smith, J. (2013). Port Resiliency Program (PReP) Final report of pilot

project at Las Americas International Airport, Santo Domingo, Dominican Republic, 5-7

February 2013. Miami, FL: AmericasRelief Team.

9. Smythe, T. C. (2013). Assessing the impacts of Hurricane Sandy on the Port of New

York and New Jersey's Maritime responders and response infrastructure. Natural Hazards

Center.

10. Rose, A., & Wei, D. (2013). Estimating the economic consequences of a port shutdown:

the special role of resilience. Economic Systems Research, 25(2), 212-232.

11. Grainger, A., & Achuthan, K. (2014). Port resilience: a primer.

12. Loh, H. S., & Thai, V. V. (2014). Managing port-related supply chain disruptions: a

conceptual paper.

13. Pant, R., Barker, K., Ramirez-Marquez, J. E., & Rocco, C. M. (2014). Stochastic

measures of resilience and their application to container terminals. Computers &

Industrial Engineering, 70, 183-194.

14. Southworth, F., Hayes, J., McLeod, S., & Strauss-Wieder, A. (2014). Making US Ports

Resilient as Part of Extended Intermodal Supply Chains (No. Project NCFRP-37).

15. Wakeman, T., Miller, J., & Python, G. (2015). Port resilience: overcoming threats to

maritime infrastructure and operations from climate change.

16. Achuthan, K., Zainudin, N., Roan, J., & Fujiyama, T. (2015). Resilience of the food

supply to port flooding on east coast.

53

17. Morris, L.L. (2016). Ports Resilience Index: Three Case Studies in the Gulf of Mexico.

Unpublished manuscript.

18. Justice, V., Bhaskar, P., Pateman, H., Cain, P., & Cahoon, S. (2016). US container port

resilience in a complex and dynamic world. Maritime Policy & Management, 43(2), 179-

191.

19. Chen, H., Cullinane, K., & Liu, N. (2017). Developing a model for measuring the

resilience of a port-hinterland container transportation network. Transportation Research

Part E: Logistics and Transportation Review, 97, 282-301.

20. Cho, H., & Park, H. (2018). Constructing Resilience Model Of Port Infrastructure Based

On System Dynamics. Disaster Management, 245.

21. Touzinsky, K. F., Scully, B. M., Mitchell, K. N., & Kress, M. M. (2018). Using

Empirical Data to Quantify Port Resilience: Hurricane Matthew and the Southeastern

Seaboard. Journal of Waterway, Port, Coastal, and Ocean Engineering, 144(4),

05018003.

22. “Marine Safety Information Bulletin (MSIB) 18-018, Hurricane Conditions, May 29,

2018”, United States Coast Guard, Sector Miami.

23. “Guidelines for the Area Maritime Security Committees and Area Maritime Security

Plans Required for U.S. Ports”, Navigation and Vessel Inspection Circular No. 9-02,

Change 4. United States Coast Guard.

24. “Guidelines for Drafting the Marine Transportation System Recovery Plan”, Navigation

and Vessel Inspection Circular 04-18. United States Coast Guard.

25. “Atlantic Area Port Operations Hurricane Guidance”, Atlantic Area Instruction

16601.1B. United States Coast Guard.

26. “United States Coast Guard Sector New York, Hurricane and Severe Weather Plan” United

States Coast Guard. May 2017.

27. Al-Deek, H. M. (2001). Comparison of Two Approaches for Modeling Freight Movement

at Seaports. Journal of Computing in Civil Engineering,15(4), 284-291.

doi:10.1061/(asce)0887-3801(2001)15:4(284)

28. Alyami, H., Yang, Z., Riahi, R., Bonsall, S., & Wang, J. (2019). Advanced uncertainty

modelling for container port risk analysis. Accident Analysis & Prevention,123, 411-421.

doi:10.1016/j.aap.2016.08.007

29. Chen, Z., Xue, J., Wu, C., Qin, L., Liu, L., & Cheng, X. (2018). Classification of vessel

motion pattern in inland waterways based on Automatic Identification System. Ocean

Engineering,161, 69-76. doi:10.1016/j.oceaneng.2018.04.072

30. Costea, A., Ticu, I. R., & Mishkoy, G. (2019). A modelling system for seaport activities.

54

31. Debnath, A. K., Chin, H. C., & Haque, M. M. (2011). Modelling Port Water Collision Risk

Using Traffic Conflicts. Journal of Navigation,64(4), 645-655.

doi:10.1017/s0373463311000257

32. Howard, D., Bragen, M., Burke, J., & Love, R. (2004). PORTSIM 5: Modeling from a

seaport level. Mathematical and Computer Modelling,39(6-8), 715-731.

doi:10.1016/s0895-7177(04)90550-x

33. John, A., Yang, Z., Riahi, R., & Wang, J. (2016). A risk assessment approach to improve

the resilience of a seaport system using Bayesian networks. Ocean Engineering,111, 136-

147. doi:10.1016/j.oceaneng.2015.10.048

34. Kozan, E., & Preston, P. (2006). Mathematical modelling of container transfers and storage

locations at seaport terminals. Container Terminals and Cargo Systems,520-537.

doi:10.1007/s00291-006-0048-1

35. Kozan, E. (2007). Comparison of analytical and simulation planning models of seaport

container terminals. Transportation Planning and Technology,20(3), 235-248.

doi:10.1080/03081069708717591

36. Magala, M., & Sammons, A. (2008). A New Approach to Port Choice Modelling. Port

Management. doi:10.1057/9781137475770.0006

37. Mangan, J., Lalwani, C., & Gardner, B. (2002). Modelling port/ferry choice in RoRo

freight transportation. International Journal of Transport Management,1(1), 15-28.

doi:10.1016/s1471-4051(01)00003-9

38. Olba, X. B., Daamen, W., Vellinga, T., & Hoogendoorn, S. P. (2018). State-of-the-art of

port simulation models for risk and capacity assessment based on the vessel navigational

behaviour through the nautical infrastructure. Journal of Traffic and Transportation

Engineering (English Edition),5(5), 335-347. doi:10.1016/j.jtte.2018.03.003

39. Tixerant, M. L., Guyader, D. L., Gourmelon, F., & Queffelec, B. (2018). How can

Automatic Identification System (AIS) data be used for maritime spatial planning? Ocean

& Coastal Management,166, 18-30. doi:10.1016/j.ocecoaman.2018.05.005

40. Resilient Interdependent Infrastructure Processes and Systems (RIPS). (2006, November

7). Retrieved from https://www.nsf.gov/pubs/2014/nsf14524/nsf14524.htm

41. U.S. Coast Guard. (2018). “U.S. Coast Guard Vessel Traffic Service User Manual.”

55

42. American Association of Port Authorities. (2015, April 21). Retrieved from

https://www.aapa-ports.org/advocating/PRDetail.aspx?ItemNumber=20424

43. Cunningham, S. (2019, May 12). Houston Ship Channel still not safe for passage to GOM

after oil spill. Retrieved from https://www.worldoil.com/news/2019/5/12/houston-ship-

channel-still-not-safe-for-passage-to-gom-after-oil-spill

44. DePillis, L. (2015, February 18). A labor dispute slowed America's ports to a halt. But

there's an even bigger problem. Retrieved from

https://www.washingtonpost.com/news/wonk/wp/2015/02/18/a-labor-dispute-slowed-

americas-ports-to-halt-but-theres-an-even-bigger-

problem/?noredirect=on&utm_term=.d4e2a3e39ab5

45. Gillespie, P., Romo, R., & Santana, M. (2017, September 28). Puerto Rico aid is trapped

in thousands of shipping containers. Retrieved from https://www.cnn.com/2017/09/27/

us/puerto-rico-aid-problem/index.html

46. NOAA. (2015, June 1). Frequently Asked Questions. Retrieved from

http://www.aoml.noaa.gov/hrd/tcfaq/E11.html

47. Resilient Interdependent Infrastructure Processes and Systems (RIPS). (2006, November

7). Retrieved from https://www.nsf.gov/pubs/2014/nsf14524/nsf14524.htm

48. Smythe, T. C., PhD. (2013, May 31). Assessing the Impacts of Hurricane Sandy on the

Port of New York and New Jersey’s Maritime Responders and Response

Infrastructure. Quick Response Report No. 238 to the University of Colorado Natural

Hazards Center. Retrieved from https://rucore.libraries.rutgers.edu/rutgers-

lib/43635/PDF/1/play/.

49. Tomer, A., & Kane, J. (2015, July 1). The top 10 metropolitan port complexes in the U.S.

Retrieved from https://www.brookings.edu/blog/the-avenue/2015/07/01/the-top-10-

metropolitan-port-complexes-in-the-u-s/#cancelNOAA. (2008, October 27). Is sea level

rising? Retrieved from https://oceanservice.noaa.gov/facts/sealevel.html

50. Welshans, Krissa. (2015, June 1). "East coast ports gaining ground: supply chain users look

to diversify to East Coast and Gulf ports after long-lasting labor dispute at West Coast

ports." Feedstuffs. Business Insights: Essentials.

51. Wolshon, B., Parr, S., Farhadi, N., & Mitchell, K. N. (2018, August 31). Quantifying and

Evaluating Community Resilience for Sustainable Cities. Maritime Transportation

Research & Education Center. Retrieved from

https://rosap.ntl.bts.gov/view/dot/37073/dot_37073_DS1.pdf?

56

52. Becker, A. H., Matson, P., Fischer, M., & Mastrandrea, M. D. (2015). Towards seaport

resilience for climate change adaptation: Stakeholder perceptions of hurricane impacts in

Gulfport (MS) and Providence (RI). Progress in Planning, 99, 1-49.

53. Southworth, F., Hayes, J., McLeod, S., & Strauss-Wieder, A. (2014). Making US Ports

Resilient as Part of Extended Intermodal Supply Chains (No. Project NCFRP-37).

54. Cho, H., & Park, H. (2017). Constructing resilience model of port infrastructure based on

system dynamics. Disaster Management, 7(3), 245-254.

55. Touzinsky, K. F., Scully, B. M., Mitchell, K. N., & Kress, M. M. (2018). Using Empirical

Data to Quantify Port Resilience: Hurricane Matthew and the Southeastern Seaboard.

Journal of Waterway, Port, Coastal, and Ocean Engineering, 144(4), 05018003.

56. Gkonis, K. G., & Psaraftis, H. N. (n.d.). Some key variables affecting liner shipping costs.

Transportation Research Board.

57

APPENDIX A – ARRIVAL AND DWELL TIME PROBABILITY

DISTRIBUTION FUNCTIONS

Containerized Cargo

Figure A1: Port of Charleston Containerized Cargo Arrival Time of Day Distributions

Figure A2: Port of Charleston Containerized Cargo Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

10.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF CHARLESTON ARRIVAL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

2%

4%

6%

8%

10%

12%

14%

16%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF CHALESTON DWELL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

58

Figure A3: Port of Savannah Containerized Cargo Arrival Time of Day Distributions

Figure A4: Port of Savannah Containerized Cargo Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF SAVANNAH ARRIVAL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF SAVANNAH DWELL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

59

Figure A5: Port of Jacksonville Containerized Cargo Arrival Time of Day Distributions

Figure A6: Port of Jacksonville Containerized Cargo Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

10.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF JACKSONVILLE ARRIVAL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF JACKSONVILLE DWELL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

60

Figure A7: Port Canaveral Containerized Cargo Arrival Time of Day Distributions

Figure A8: Port Canaveral Containerized Cargo Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%

16.00%

18.00%

20.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT CANAVERAL ARRIVAL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

5%

10%

15%

20%

25%

30%

35%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT CANAVERAL DWELL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

61

Figure A9: Port Everglades Containerized Cargo Arrival Time of Day Distributions

Figure A10: Port Everglades Containerized Cargo Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT EVERGLADES ARRIVAL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT EVERGLADES DWELL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

62

Figure A11: Port of Miami Containerized Cargo Arrival Time of Day Distributions

Figure A12: Port of Miami Containerized Cargo Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%

16.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF MIAMI ARRIVAL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF MIAMI DWELL TIME DISTRIBUTIONS: CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

63

Non-Containerized Cargo

Figure A13: Port of Charleston Non-Containerized Cargo Arrival Time of Day Distributions

Figure A14: Port of Charleston Non-Containerized Cargo Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

10.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF CHARLESTON ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

1 4 710 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97

100

103

106

109

112

115

118

121

124

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF CHALESTON DWELL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

64

Figure A15: Port of Savannah Non-Containerized Cargo Arrival Time of Day Distributions

Figure 12: Port of Savannah Non-Containerized Cargo Dwell Time Distributions

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF SAVANNAH ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

1%

2%

2%

3%

3%

4%

4%

5%

5%

1 4 710 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97

100

103

106

109

112

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF SAVANNAH DWELL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

65

Figure A17: Port of Jacksonville Non-Containerized Cargo Arrival Time of Day Distributions

Figure A18: Port of Jacksonville Non-Containerized Cargo Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF JACKSONVILLE ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF JACKSONVILLE DWELL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

66

Figure A19: Port Canaveral Non-Containerized Cargo Arrival Time of Day Distributions

Figure A20: Port Canaveral Non-Containerized Cargo Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT CANAVERAL ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

2%

4%

6%

8%

10%

12%

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT CANAVERAL DWELL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

67

Figure A21: Port Everglades Non-Containerized Cargo Arrival Time of Day Distributions

Figure A22: Port Everglades Non-Containerized Cargo Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT EVERGLADES ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

10%

20%

30%

40%

50%

60%

70%

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT EVERGLADES DWELL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

68

Figure A23: Port of Miami Non-Containerized Cargo Arrival Time of Day Distributions

Figure A24: Port of Miami Non-Containerized Cargo Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF MIAMI ARRIVAL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

1 4 710 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97

100

103

106

109

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF MIAMI DWELL TIME DISTRIBUTIONS: NON-CONTAINERIZED CARGO VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

69

Tanker Vessels

Figure A25: Port of Charleston Tanker Vessel Arrival Time of Day Distributions

Figure A26: Port of Charleston Tanker Vessel Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF CHARLESTON ARRIVAL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

2%

4%

6%

8%

10%

12%

14%

16%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF CHALESTON DWELL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

70

Figure A27: Port of Savannah Tanker Vessel Arrival Time of Day Distributions

Figure A28: Port of Savannah Tanker Vessel Dwell Time Distributions

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF SAVANNAH ARRIVAL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF SAVANNAH DWELL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

71

Figure A29: Port of Jacksonville Tanker Vessel Arrival Time of Day Distributions

Figure A30: Port of Jacksonville Tanker Vessel Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF JACKSONVILLE ARRIVAL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

2%

4%

6%

8%

10%

12%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF JACKSONVILLE DWELL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

72

Figure A31: Port Canaveral Tanker Vessel Arrival Time of Day Distributions

Figure A32: Port Canaveral Tanker Vessel Dwell Time Distributions

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%

16.00%

18.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT CANAVERAL ARRIVAL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

2%

4%

6%

8%

10%

12%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT CANAVERAL DWELL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

73

Figure A33: Port Everglades Tanker Vessel Arrival Time of Day Distributions

Figure A34: Port Everglades Tanker Vessel Dwell Time Distributions

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT EVERGLADES ARRIVAL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT EVERGLADES DWELL TIME DISTRIBUTIONS:TANKER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

74

Passenger Vessels

Figure A35: Port of Charleston Passenger Vessel Arrival Time of Day Distributions

Figure A36: Port of Charleston Passenger Vessel Dwell Time Distributions

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF CHARLESTON ARRIVAL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

10%

20%

30%

40%

50%

60%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF CHALESTON DWELL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

75

Figure A37: Port of Charleston Passenger Vessel Arrival Time of Day Distributions

Figure A38: Port of Savannah Passenger Vessel Dwell Time Distributions

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

30.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF SAVANNAH ARRIVAL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

5%

10%

15%

20%

25%

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF SAVANNAH DWELL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

76

Figure A39: Port of Jacksonville Passenger Vessel Arrival Time of Day Distributions

Figure A40: Port of Jacksonville Passenger Vessel Dwell Time Distributions

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

30.00%

35.00%

40.00%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF JACKSONVILLE ARRIVAL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

5%

10%

15%

20%

25%

30%

35%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF JACKSONVILLE DWELL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

77

Figure A41: Port Canaveral Passenger Vessel Arrival Time of Day Distributions

Figure A42: Port Canaveral Passenger Vessel Dwell Time Distributions

0%

2%

4%

6%

8%

10%

12%

14%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT CANAVERAL ARRIVAL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

5%

10%

15%

20%

25%

30%

35%

40%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT CANAVERAL DWELL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

78

Figure A43: Port Everglades Passenger Vessel Arrival Time of Day Distributions

Figure A44: Port Everglades Passenger Vessel Dwell Time Distributions

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT EVERGLADES ARRIVAL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT EVERGLADES DWELL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

79

Figure A45: Port of Miami Passenger Vessel Arrival Time of Day Distributions

Figure A46: Port of Miami Passenger Vessel Dwell Time Distributions

0%

5%

10%

15%

20%

25%

30%

35%

40%

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

ARRIVAL BINS

PORT OF MIAMI ARRIVAL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.

0%

5%

10%

15%

20%

25%

30%

1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031323334353637383940414243

PR

OB

AB

ILIT

Y D

ISTR

IBU

TIO

N

DWELL TIME

PORT OF MIAMI DWELL TIME DISTRIBUTIONS: PASSENGER VESSELS

12 MONTH CAL. 10 YEAR SIM. 30 DAY SIM. 3 MONTH VAL.